deterministic-apt-layer
318 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Eric Allam
|
90e8bd5c12 |
feat(webapp,database): opt-in per-client Prisma driver adapters (#4539)
🚀 Publish Trigger.dev Docker / typecheck (push) Failing after 0s
🚀 Publish Trigger.dev Docker / units (push) Failing after 0s
🚀 Publish Trigger.dev Docker / publish-webapp (push) Has been skipped
🚀 Publish Trigger.dev Docker / publish-worker-v4 (push) Has been skipped
🚀 Publish Trigger.dev Docker / scan-webapp (push) Has been skipped
🚀 Publish Trigger.dev Docker / scan-supervisor (push) Has been skipped
🦋 Changesets PR / Create Release PR (push) Has been cancelled
🚀 Publish Trigger.dev Docker / 📣 Dispatch main image (push) Has been cancelled
📚 Docs Checks / check-broken-links (push) Has been cancelled
🧭 Helm Chart Prerelease / lint-and-test (push) Has been cancelled
Workflow Checks / Actionlint (push) Has been cancelled
Workflow Checks / Zizmor (push) Has been cancelled
🧭 Helm Chart Prerelease / prerelease (push) Has been cancelled
## What
Adds an opt-in path to run each Prisma client through
**`@prisma/adapter-pg`** (the node-postgres driver) instead of the
built-in engine driver, controlled by a **per-client env var, all off by
default**:
| env var | client |
|---|---|
| `CONTROL_PLANE_DATABASE_WRITER_DRIVER_ADAPTER` | control-plane writer
|
| `CONTROL_PLANE_DATABASE_REPLICA_DRIVER_ADAPTER` | control-plane
replica |
| `RUN_OPS_DATABASE_WRITER_DRIVER_ADAPTER` | new run-ops writer |
| `RUN_OPS_DATABASE_REPLICA_DRIVER_ADAPTER` | new run-ops replica |
| `RUN_OPS_LEGACY_DATABASE_WRITER_DRIVER_ADAPTER` | legacy run-ops
writer |
| `RUN_OPS_LEGACY_DATABASE_REPLICA_DRIVER_ADAPTER` | legacy run-ops
replica |
With every flag unset the construction path is byte-identical to today
(`datasources` URL + Rust engine), so this is inert until a flag is
turned on. Per-client granularity allows enabling the adapter only where
it's wanted.
## How
- Enables the `driverAdapters` preview feature on both schemas
(`@trigger.dev/database` and `@internal/run-ops-database`). This keeps
the **Rust query engine** — it does NOT add `queryCompiler` — so query
behavior, result types, and engine tracing spans are unchanged.
- A shared `buildDriverAdapterPool` builds each client's `pg.Pool` with
an explicit `max`, a bounded `connectionTimeoutMillis` (the
node-postgres pool otherwise waits unbounded on acquire), and an
`onPoolError` handler (an unhandled idle-connection error would
otherwise crash the process). Threaded through all four client builders
via a `useDriverAdapter` flag.
- Adds `@prisma/adapter-pg` + `@types/pg` to the webapp; `pg` is already
pinned at `8.15.6` (adapter-pg 6.x requires `pg < 8.17`).
## Connect-failure handling (the important correctness/security bit)
Under the adapter an unreachable DB no longer surfaces as
`PrismaClientInitializationError` / `P1001`; it becomes a `P2010`
"Database not reachable: <host>" (or a raw
`ECONNREFUSED`/`ENOTFOUND`-class error). Two handlers are updated so a
client on the adapter behaves like today:
- **`isInfrastructureError`** now recognizes those shapes (P2010 with a
connectivity message, and raw connectivity errno codes). Without this,
the DB **hostname would leak into API-client-facing errors** and the
failure would go unlogged. Security-relevant.
- **`isPrismaRetriableError`** treats the adapter's pool-acquire timeout
("timeout exceeded when trying to connect") as retriable, preserving the
`P2024` retry behavior the adapter otherwise drops.
## Evidence
Validated on an isolated stack that mirrors the production DB topology
(chained PgBouncers in front of writer + reader):
- **Behavioral parity:** raw-query results and Prisma error codes/`meta`
are byte-identical between the engine driver and the adapter across the
queried shapes (unique-constraint `meta.target`, record-not-found,
transaction-timeout, serialization-failure, etc.).
- **Feature matrix:** a full 380-project queue-ay pass shows no
adapter-caused regressions — pass/fail parity between adapter-off and
adapter-on, with the residual failures being pre-existing
known-failures/flakes common to both.
## Rollout / rollback
All flags default off; enable per client via env var, roll back by
unsetting and redeploying (no data migration). Recommended first target
is a single writer; enable one client at a time.
## Follow-ups (not in this PR)
- `$metrics`-based pool observability is removed under the adapter (the
Prometheus route + `db.pool.connections.*` instruments); the metrics
replacement (via `pg.Pool` counters) lands in a separate PR.
- Note for operators: on the adapter path, interactive-transaction
`maxWait` does not bound pool acquisition — `connectionTimeoutMillis`
does.
## Note on connection-string parameters
The adapter pool is built from the base DSN, so Prisma-specific DSN
parameters that node-postgres does not understand are not honored when a
client is on the adapter:
- **Prisma TLS spellings** (`sslaccept`, `sslcert`, etc.) —
node-postgres uses `sslmode`/`ssl` instead. Our production DSNs do not
use these Prisma-specific TLS params, but any deployment whose DSN
relies on them must be checked before enabling a flag.
- `pgbouncer=true` and `statement_cache_size` — effectively moot under
the adapter, which uses no persistent named prepared statements.
`connection_limit`, `pool_timeout`, and `schema` are handled explicitly
(passed as `max`/`connectionTimeoutMillis` and PrismaPg's `{schema}`
option).
refs TRI-13039
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|
Matt Aitken
|
04f9c4e1a5 |
fix(webapp,run-engine,core): drop the hidden debounce ceiling, fail fast on an unusable maxDelay (#4521)
Debouncing with a `delay` longer than an hour did nothing at all. The engine applied a server-side ceiling on how long a debounced run could be pushed back, measured from the run's `createdAt` and defaulting to one hour. A run is only pushed back while its new execution time stays inside that ceiling, so a `delay` at or above it could never push anything: the waiting run was released, the trigger started its own run, and the next trigger repeated it. A `delay: "12h"` produced one run per trigger, each correctly delayed by 12h, with no error raised and nothing on the run to show the debounce key had been ignored. The ceiling is now unset by default. A debounce key with no `maxDelay` keeps collapsing triggers for as long as they keep arriving, which is what the docs have always described. Self-hosters who want a bound can still set `RUN_ENGINE_MAXIMUM_DEBOUNCE_DURATION_MS`. That has a consequence worth stating plainly, so the docs now carry a warning for it: with no `maxDelay`, a continuously triggered key never executes. Set `maxDelay` when the work has to happen eventually. **Failing fast on an unusable `maxDelay`.** A caller who sets `maxDelay` no longer than their `delay` hits exactly the dead end described above, so that pair is now rejected at trigger time instead of silently behaving as if no debounce were set: ``` debounce.maxDelay (1h) must be longer than debounce.delay (12h). A debounced run is only pushed back while it stays inside maxDelay, so with these values every trigger would create its own run. ``` An unparseable `maxDelay` is rejected too, rather than quietly falling back to no bound at all, and so is a `delay` given as a date rather than a duration, which could never work because the value is re-applied on every push. The same check runs against a configured server ceiling, so a self-hosted deployment that sets `RUN_ENGINE_MAXIMUM_DEBOUNCE_DURATION_MS` gets the error rather than the silent failure this PR is about. With no `maxDelay` and no configured ceiling, which is the default, there is nothing to conflict with and nothing is rejected. The docs, the `TriggerOptions` JSDoc and the engine option all now state that the room available to push is the gap between `delay` and `maxDelay`. The run engine suite gains the case that motivated this: four triggers on one key with a 12h delay now collapse to a single run. |
||
|
Eric Allam
|
58bf4e2833 |
feat(webapp): per-client database pool and connect timeout overrides (#4515)
## Summary Follow-on to #4513. The database connect timeout is now honored, but a single global value has to serve three separate databases at once (control-plane, legacy run-ops, and run-ops). This adds optional per-client overrides for the Prisma pool and connect timeouts, one pair for the writer and one for the read replica of each of the three databases, each falling back to the shared `DATABASE_POOL_TIMEOUT` / `DATABASE_CONNECTION_TIMEOUT` when unset. That lets one database's clients run a fail-fast connect timeout (with a bounded pool wait) while another keeps more headroom, without a single knob forcing the same tradeoff everywhere. No behavior change until an override is set. It also tags each client's queries with its specific datasource (`control-plane` / `legacy-run-ops` / `run-ops`, writer or replica) via the `db.datasource` span attribute, so telemetry can attribute connection behavior to a specific database instead of just writer-vs-replica. |
||
|
Eric Allam
|
db6228dd1e | chore(webapp,core,sdk): upgrade @s2-dev/streamstore to 0.25 and migrate S2 hosts (#4349) | ||
|
Eric Allam
|
0445b8ec27 |
fix(webapp,clickhouse): keep the rest of a ClickHouse batch when one run or span has un-ingestable JSON (#4358)
## Summary A single run output, trace span, or payload carrying JSON that ClickHouse can't ingest (for example nesting past its depth limit) used to fail the whole insert batch, so unrelated runs and spans silently disappeared from the runs list, traces, and logs. This keeps the rest of the batch and handles the offending row instead of dropping everything around it. ## Fix Recovery is per-table, matched to what each table needs: - **Runs** (`task_runs_v2`) keep their status. We follow ClickHouse's failing-row hint to strip just the un-ingestable JSON column(s) so the run still lands (its output reads from Postgres on the detail page), up to a configurable limit (`RUN_REPLICATION_MAX_POISON_STRIPS_PER_BATCH`, default `1`). Past the limit we stop and land the batch with `allow_errors` in a single pass, skipping the remainder. Cost stays a fixed handful of inserts no matter how large or poisoned a flush is. - **Trace events and payloads** (high volume, append-only) recover with a single `allow_errors` insert: the good rows land in one pass and only the un-ingestable rows are skipped. Before falling back, a lightweight sanitizer still repairs what it can losslessly (lone UTF-16 surrogates, out-of-range integers) so a repairable row lands in full. To read the failing-row hint we patch `@clickhouse/client-common`: its error parser truncates the server response and discards the `(at row N)` position, so the patch preserves the full text for the recovery path to read. |
||
|
Eric Allam
|
c72ebf9084 |
fix(webapp,run-engine): stop batchTriggerAndWait hanging when item streaming never completes (#4397)
## Summary `batchTriggerAndWait()` could leave a parent run waiting forever. The 2-phase batch API blocks the parent on the batch's waitpoint as soon as the batch is created, but the batch is only sealed at the end of item streaming. If streaming never completed, nothing sealed the batch, nothing completed the waitpoint, and the parent stayed suspended with no timeout and no way to recover. Supersedes #4016, which added the reaper alone. ## Fix Admission for item streaming was being decided twice. Batch creation passes its own rate limiter, which fixes `expectedCount` and blocks the parent, and then the item stream had to pass the general API limiter as well, competing with unrelated traffic. A second limiter could therefore veto work the first had already committed the parent to. Creation now mints a bounded grant that the item stream spends, so an admitted batch can finish streaming. The grant is capped per batch rather than exempting the path, and every failure mode (no grant, spent grant, unreachable store) falls back to the normal limiter. That makes stranding much rarer but not impossible, since a request timeout or a crash can still end streaming for good. So a seal-timeout reaper aborts any batch still unsealed after `BATCH_SEAL_TIMEOUT_MS` and completes the parent's waitpoint with an error, letting `batchTriggerAndWait()` reject instead of hang. It is race-safe against a late seal, and it is only scheduled for batches that actually block a parent, so fire-and-forget batches cost nothing. Finally, the batches page used to report "Batch completion checked." for these batches while doing nothing, because the completion path returns early on an unsealed batch. It now says the batch cannot be resumed. Rate limiting is no longer the reason a batch strands, so the reaper's default stays at 30 minutes, comfortably above the SDK's worst-case stream-retry budget. ## Verification Unit and container tests cover the grant cap, the bypass ordering (it runs after the authorization check, so it can never skip authentication), and the reaper's abort, seal race, idempotency, and no-waitpoint cases. Also verified end-to-end against a running stack. With the general limit exhausted, batch creation and other API calls returned 429 while a granted batch still streamed and sealed; an ungranted batch id was rate limited rather than bypassed; and the grant cut off exactly at its configured attempt count. Reproducing the stranded state on a real parent run, the batch was aborted at the timeout, the waitpoint completed with an error, and the parent resumed and finished instead of hanging. A parentless batch left unsealed was untouched well past the reaper window. ## Verified against deployed runs The reaper was proven end to end with a real deployed run (locally-run supervisor, containerised run) and a real network fault, rather than a simulated one: toxiproxy severs the phase 2 item stream mid-flight so every SDK stream retry genuinely fails, while phase 1 still succeeds. Only the batch calls traverse the fault, so control-plane traffic is untouched. The reproduction is the shape that actually strands a parent: the task catches the `BatchTriggerError` the SDK throws and carries on, so the phase 1 block outlives the thrown error and the parent hangs at its next suspension point. With the reaper disabled, the parent sat in `EXECUTING_WITH_WAITPOINTS` for over 24 minutes holding two blockers, and stayed stuck across a full infrastructure restart: ``` type | status | has_timeout BATCH | PENDING | f <- orphan, completedAfter NULL DATETIME | COMPLETED | t <- the wait already elapsed ``` With the reaper enabled the same task under the same fault completed in about 75 seconds with zero blockers left, the batch `ABORTED`, and its waitpoint completed carrying the error. Two conditions are required to observe this at all, which is worth knowing for any future test: the run must be deployed rather than `trigger dev` (dev runs execute in process and finish while still holding blocker rows), and the wait after the caught error must exceed the checkpoint threshold, or it is served in process and never suspends. ### Why completing the batch waitpoint is sufficient `batchTriggerAndWait` runs create, then stream, then wait. A phase 2 failure throws before the wait is ever reached, and the reaper only fires on an unsealed batch, so the parent is never suspended awaiting the batch when it runs. The parent therefore does not need a synthetic result, only to stop being blocked. Note this reasoning depends on that ordering: if the wait were ever reached with an unsealed batch, completing the batch waitpoint alone would not settle the caller. ## Follow-ups - Batches stranded before this ships still need a one-off recovery; the reaper only schedules at creation time. - That same property leaves a gap if the process dies between creating the batch and scheduling the job. A periodic sweep would close it, but wants a supporting index. - When a partially streamed batch aborts, children already enqueued keep running while the parent fails. Left as-is deliberately, since cancelling triggered work is a bigger semantic call. |
||
|
Eric Allam
|
4eb9292cbe |
feat(webapp,run-engine): queue metrics and health dashboard (#4131)
## Summary
Three related changes, each independently gated:
**Queue metrics and health.** Per-queue depth, throughput (enqueued,
started, completed), concurrency, whether a queue is throttled, and
scheduling delay (how long a run waits between becoming eligible and
actually starting), plus a per concurrency-key breakdown for keyed
queues. Collected from inside the run queue itself, stored in
ClickHouse, and surfaced on the Queues list, a new per-queue detail
page, the task pages, and the run inspector. The question it answers is
"does this queue have enough concurrency to keep up, and if not, which
key or which limit is the constraint".
**Percent-based queue concurrency limits.** A queue's concurrency
override can now be expressed as a percentage of the environment limit,
stored as the source of truth and re-materialized whenever the
environment limit changes. Absolute overrides above the environment
limit are now **rejected with a 400** instead of being silently capped,
which is a behavior change on `POST
/api/v1/queues/:queue/concurrency/override`.
**The `health` report.** A server-computed verdict on whether work is
flowing, whether the runs that do start are healthy, and whether
telemetry is fresh, rendered as text with sparklines. Available as `GET
/api/v1/reports/:key`, `trigger report`, and the `get_report` MCP tool
(plus a `report` MCP prompt, which shows up as a slash command in hosts
that support prompts).
With the flags off, the Queues page renders the pre-metrics component
verbatim, nothing is emitted, and nothing is written to ClickHouse.
## Configuration
Two independent gates, on purpose. Emission is global so data accrues
for everyone before anyone can look at it; the view is per organization
so it can be turned on for one org at a time without a deploy.
**Runtime flags (no restart)**
| Flag | Store | Gates |
| --- | --- | --- |
| `queue_metrics:enabled` | run-queue Redis key (`"1"`/`"0"`, off by
default) | All emission, gauges and counters. Cached in-process for 10s
with stale-while-revalidate, warmed eagerly at boot so the first op
after a deploy is not dropped. |
| `queue_metrics:gauge_sample_rate` | run-queue Redis key, `0..1` |
Fraction of queue ops that emit a gauge. Counters are never sampled, so
throughput stays exact at any rate. |
| `queueMetricsUiEnabled` | feature-flag catalog: global `FeatureFlag`
row, per-org `Organization.featureFlags` override wins | Whether an org
sees the metrics view at all: the Queues list variant, the queue detail
route, the built-in Queues dashboard, the concurrency-keys endpoint, and
the metrics blocks on task pages and the run inspector. Off by default;
a gated org gets a 404 on the detail route rather than an empty page. |
Both Redis keys are readable and writable from `/admin/queue-metrics`
(super-admin UI, with a live per-shard stream-health table) and
`GET`/`POST /admin/api/v1/queue-metrics` (admin PAT). The admin surface
uses its own Redis client, so it works on any instance regardless of
whether that instance runs the emitter or the consumer.
**Environment variables (boot time)**
| Variable | Default | Notes |
| --- | --- | --- |
| `QUEUE_METRICS_EMIT_ENABLED` | `0` | Constructs the emitter and
injects it into the run engine. Without it the run queue has no emitter
at all. |
| `QUEUE_METRICS_CONSUMER_ENABLED` | `0` | Boots the stream consumer on
this instance. Independent of emission, so consumers can be sized
separately from the API. |
| `QUEUE_METRICS_STREAM_SHARD_COUNT` | `4` | Stream shards, hashed per
queue. |
| `QUEUE_METRICS_CONSUMER_BATCH_SIZE` | `1000` | Poll batch equals
insert batch, so an ack can never outrun a write. |
| `QUEUE_METRICS_REDIS_{HOST,PORT,USERNAME,PASSWORD,TLS_DISABLED}` |
falls back to the run-queue Redis | Set `HOST` to move the metrics
stream onto a dedicated instance so a metrics backlog cannot compete
with the run queue for memory. Self-hosters can leave it unset and get a
single-Redis deployment. |
| `QUEUE_METRICS_COUNTER_STREAM_MAXLEN` | `2000000` shared, `8000000`
dedicated | Bound on how much a stalled consumer can hold. The default
is deliberately lower when the stream shares the queue-critical Redis. |
| `QUEUE_METRICS_COUNTER_ODOMETER_TTL_SECONDS` | `604800` | TTL on the
per-queue cumulative counter key, refreshed on every write, so only
queues idle for the whole window are purged. |
| `QUEUE_METRICS_MAX_QUEUE_NAMES_PER_ENV` | `1000` | Distinct queue
names tracked per environment; overflow collapses into `__overflow__`. |
| `QUEUE_METRICS_MAX_CONCURRENCY_KEYS_PER_QUEUE` | `10000` | Same idea
one level down, per queue. |
| `QUEUE_METRICS_GAUGE_SAMPLE_RATE` | `1` | Default for the live
sample-rate key above. |
| `QUEUE_METRICS_QUERY_TABLES_VISIBLE` | `0` | Lists the queue-metrics
tables in the Query page, its schema docs, the schema API and the AI
query context. Off keeps them unlisted while the feature is dark; a
query naming them still runs either way. |
| `QUEUE_METRICS_CLICKHOUSE_URL` | falls back to the shared wiring |
Runs queue metrics on their own ClickHouse service: the consumer's
inserts and every queue-metrics read go through it, so a metrics-heavy
chart refresh never competes with runs-list or trace reads. Unset
reproduces the previous split exactly (inserts on `CLICKHOUSE_URL`,
reads on the query pool). |
| `QUEUE_METRICS_CLICKHOUSE_READER_URL` | the write URL | Reader split,
so the consumer's inserts can never land on a read endpoint. |
|
`QUEUE_METRICS_CLICKHOUSE_{KEEP_ALIVE_ENABLED,KEEP_ALIVE_IDLE_SOCKET_TTL_MS,MAX_OPEN_CONNECTIONS,LOG_LEVEL,COMPRESSION_REQUEST}`
| `1`, unset, `10`, `info`, `1` | Pool tuning, matching the other
per-workload ClickHouse clients. |
Migrations to apply: ClickHouse `036_create_queue_metrics_v1.sql`, and a
Postgres migration adding the nullable
`TaskQueue.concurrencyLimitOverridePercent`. Both are additive.
## How collection works
Queue operations produce two kinds of signal, and they have opposite
failure modes, so they are handled differently.
**Gauges** (queued, running, queue limit, env queued, env running, env
limit, throttled, plus keys-with-backlog and worst-key wait on keyed
queues) are read *inside* the same Redis script that performs the
enqueue or dequeue, so the reading is atomic with the operation it
describes rather than a racy follow-up read. The script returns them on
its reply and the app forwards them to the stream. Gauges are sampled
and drop-tolerant: they are aggregated with `max`, so a lost reading
costs resolution, never correctness.
**Counters** (enqueued, started, completed, plus nack and dead-lettered)
are cumulative odometers. Each event increments a per-queue key on the
metrics Redis and emits the absolute total, and ClickHouse takes the
difference across buckets at read time. This is the important property
of the design: a summed-delta counter undercounts permanently on any
lost event, while a cumulative one self-heals, because the next
surviving reading restates the whole total. Only bucket granularity can
be lost, never the total. A queue returning after its odometer TTL
expired restarts at 1 and reset detection handles it, which is safe
precisely because expiry only spans a window with no activity.
Both land on one sharded Redis stream. A consumer reads it with a
consumer group, reclaims stale pending entries on a 15s interval rather
than on every poll, maps one entry to one or two ClickHouse rows
(whole-queue and, for keyed queues, per-key), and acks only after the
insert lands. Each batch carries a dedup token derived from its
stream-entry ids, and the target tables set
`non_replicated_deduplication_window`, so a retried batch cannot
double-count either the raw rows or the aggregates that hang off them.
Consumer and emitter both emit OTel metrics
(`queue_metrics.emitter.emitted`,
`queue_metrics.consumer.{entries,rows_inserted,insert_errors,insert_duration,stream_depth,group_lag,pending,lag_unknown}`);
stream depth and group lag are the two worth alerting on, and
`lag_unknown` exists because Redis can report a null lag after a trim,
which must not be read as zero.
## Storage and read path
`queue_metrics_raw_v1` is a short landing table with a 6 hour TTL. Four
aggregate tiers are materialized straight from raw, never cascaded off
each other, each with a 30 day TTL:
- `queue_metrics_v1`, 10 second buckets per queue, the default read path
- `queue_metrics_5m_v1`, 5 minute buckets per queue, for wide ranges and
cross-queue ranking
- `env_metrics_v1`, 10 second buckets per environment, queue-independent
so it stays cheap at any range
- `queue_metrics_ck_v1`, 10 second buckets per concurrency key
Every tier is an MV from raw because the counter states do not survive a
cascade: their merge is order sensitive, so a `-MergeState` chain off
the 10s table inflates the result, and the same property means an
aggregate state may only be merged inside one queue. That constraint is
now enforced by the query engine rather than by reviewer discipline: a
column can declare a `mergeGroupKey`, and any query that references it
without grouping by, or pinning to a single value of, every named key
fails to compile with an actionable message.
On the read side, TRQL gains three tables (`queue_metrics`,
`env_metrics`, and a `queue_metrics_by_key` that is hidden from the
editor, schema docs and schema API but still queryable, so per-key rows
can never silently merge into a plain per-queue query), plus
`deltaSumTimestampMerge` and `quantilesTDigestMerge`. Two schema-level
optimizations ride along: a table can declare coarser rollups, so a
query whose bucket interval is 5 minutes or wider is routed to the 5m
table with no change to the query itself, and it can opt into the
ClickHouse query cache with time bounds floored to a fixed grid, so the
auto-refreshing dashboards actually share cache entries instead of
missing on every tick. Both are caller-side substitutions, so the
printer stays unaware of physical layout.
All of this can also live on its own ClickHouse service. A table
declares the pool its reads run on, the three queue-metrics tables name
the dedicated one, and the ingestion consumer writes through the same
client, so both directions move together with one env var and nothing
else routes differently.
The other engine change is opt-in gap filling: charts can request rows
for empty buckets, where counters zero-fill and gauges carry forward.
Grouped gauge series are densified per group and carried inside a
partition, so a quiet queue's line holds its last value without bleeding
another queue's value into it.
## Queue concurrency limits
`concurrencyLimitOverridePercent` on `TaskQueue` is the source of truth
when an override is set as a percentage; the absolute `concurrencyLimit`
is materialized from it (floored, clamped to at least 1 so a percentage
can never act as a pause, and never above the environment limit). Every
path that changes an environment limit now recalculates the
environment's percent-based overrides afterwards, outside the
transaction, and pushes changed limits to the engine. The push is
attempted even when the stored value did not change, so a previously
failed sync self-heals rather than leaving the database and the engine
diverged; paused queues are skipped so a recalculation cannot
effectively unpause one.
The API accepts exactly one of `concurrencyLimit` or `percent`, and the
reject-instead-of-clamp change above means a request asking for more
than the environment allows now fails loudly. The percent bound (greater
than 0, at most 100) is defined once and shared by the zod schema, the
dashboard mutation handler and the service, so the three cannot drift.
The concurrency-keys table on a queue is now paginated against the
ClickHouse per-key tier, ranked by peak backlog with the total on every
row from a single scan, and only the keys on the current page are
enriched with live counts from Redis. That replaces a hard top-50 cap
with something whose cost is a function of page size rather than key
cardinality.
## The health report
`GET /api/v1/reports/:key?period=&format=markdown|ansi|json`. The
verdict is computed on the server and is deterministic, not
model-generated. Three independent analyzers run over one input
snapshot: flow (is work moving, and if not, is the cause a limit,
throttling, one bad queue, or dead-lettering), execution (are the runs
that start succeeding, and at what latency), and liveness (how fresh is
the telemetry). When telemetry is genuinely stale, the first two are
forced to unknown and every actionable field is stripped, so no surface
ever advises action off stale data.
Authorization is per query table rather than a blanket query grant: a
JWT must be scoped to every table the report reads (`runs`,
`env_metrics`, `queue_metrics`), so a narrowly scoped token cannot pull
a report that reads more than it was granted. `period` is validated as a
shorthand with a 90 day ceiling at the edge. The report catalog is a
registry of `{ load, interpret }` entries, so the next report is a new
entry and no change to the route, the view model, the renderers, the CLI
or the MCP tool.
`trigger mcp` no longer launches the install wizard when stdout is a
TTY, which fixed a real failure: hosts spawn the server over a PTY, so
the wizard would open and the client would time out waiting for a server
that never started. The wizard now needs `trigger mcp --install`.
## The part that is live regardless of every flag
The enqueue and dequeue scripts now return a 2-tuple so a gauge reading
can ride back on the reply. Every return site in the eight affected
scripts is wrapped, and a `nil` original is converted to `false` on the
way out, because a raw `nil` in the first slot would make Lua truncate
the multi-bulk reply and silently drop the gauge on the throttled and
empty-queue paths. The reply shape and the destructuring on the app side
are exercised on every queue operation whether or not metrics are
enabled, so that is the part of `run-engine` worth the closest review.
One behavior fix in the same area: the scheduling-delay anchor is set
only on a run's first entry into the queue. Anchoring it to trigger time
on re-enqueues made waitpoint and checkpoint resumes report the entire
wait as scheduling delay. Queue ordering is untouched, so a re-enqueued
run keeps its position, and nacks deliberately keep the original anchor
because a rolled-back dequeue is the same continuous wait.
A pending-version promotion still anchors to trigger time, on purpose:
that promotion is the run's first real entry into the queue, since the
trigger deliberately held it back waiting for a worker version, and the
TTL is armed at the same point for the same reason. The consequence is
worth naming, because it is a judgement call: a run that waits on a
deployment reports that wait as scheduling delay on its queue, which is
time unrelated to queue capacity.
## Verification
Unit and integration suites across the new package, the run queue, the
mapping layer, the query engine and ClickHouse (including a test that
applies migration 036 through the same splitter CI uses, and a
regression test that inserts the same batch three times to prove the
aggregates do not inflate). Beyond that, the whole path was driven end
to end against a live stack with real runs: emitter to Redis stream to
consumer to ClickHouse to the dashboards, for both the local dev path
and the deployed path where a supervisor drives the dequeue, with
assertions on exact counter reconstruction per queue and per concurrency
key, throttling, environment saturation, scheduling delay, and a
deliberate mid-stream reading drop to confirm the cumulative counters
still reconstruct the correct total. The gated-off state was checked on
every touched surface.
The dedicated ClickHouse service was verified against a second,
separately-schema'd instance: with it configured, the driven counters
reconstruct exactly on the dedicated instance, the shared instance gains
no rows for that window, a read through the query API returns the value
that exists only on the dedicated instance, and a `runs` query still
succeeds (it would fail outright if it were mis-routed to a service
without that table). With the variable unset, the full suite passes
unchanged.
---------
Co-authored-by: Katia Bulatova <katia@trigger.dev>
Co-authored-by: Katia Bulatova <katherine.bulatova@gmail.com>
Co-authored-by: James Ritchie <james@trigger.dev>
|
||
|
Saadi Myftija
|
44eca4d166 |
feat(webapp): org-gated internal API origin in run env vars (#4366)
Adds an opt-in way for operators to route deployed runs' API traffic through a different origin than the public one, per organization. Set `INTERNAL_API_ORIGIN` on the webapp and enable the `internalApiOriginEnabled` feature flag (globally or per org, with the org override winning in both directions): deployed runs for enabled orgs then get `TRIGGER_API_URL` set to the internal origin instead of `API_ORIGIN`. Useful for gradually moving run traffic onto a private network path. ## Design The origin is resolved when an attempt starts, so flag changes take effect on the next attempt and roll back the same way, with no task redeploys. The org override is read fresh per attempt; the global default comes from the cached flags registry (a cold read fails safe to the public origin). When `INTERNAL_API_ORIGIN` is unset the flag is a no-op and no extra queries run, so existing deployments are unaffected. Dev runs always use the public origin, and `TRIGGER_STREAM_URL` remains unchanged. |
||
|
Eric Allam
|
d3906241a5 |
feat(webapp): read realtime run rows from the primary, not the replica (#4378)
🦋 Changesets PR / Create Release PR (push) Has been cancelled
🚀 Publish Trigger.dev Docker / 📣 Dispatch main image (push) Has been cancelled
🚀 Publish Trigger.dev Docker / typecheck (push) Failing after 2s
🚀 Publish Trigger.dev Docker / units (push) Failing after 2s
🚀 Publish Trigger.dev Docker / publish-webapp (push) Has been skipped
🚀 Publish Trigger.dev Docker / publish-worker-v4 (push) Has been skipped
🚀 Publish Trigger.dev Docker / scan-webapp (push) Has been skipped
🚀 Publish Trigger.dev Docker / scan-supervisor (push) Has been skipped
## Summary
The realtime runs feed hydrates run rows from read replicas, which means
it needs a replica-lag gate to avoid serving a run's previous state
right after a write. Setting
`REALTIME_BACKEND_NATIVE_RUN_READS_FROM_PRIMARY=1` reads those rows from
each run store's primary instead, so there is no lag to gate against: no
probe, no wake delay, no stale-read retries. Off by default, so nothing
changes unless you set it.
## Design
The run stores already decide replica-vs-primary from the *brand* on the
read client they are handed: a branded replica keeps the read on the
owning store's replica, an unbranded writer escalates it to that store's
own primary. So this is a one-line choice at the hydrator, and it stays
correct across topologies. With the run-ops split on, each leg lands on
its own writer and the caller's client is never forwarded across
databases; with the split off, it is the single database's primary.
```ts
const runReader = new RunHydrator({
readClient: runReadsFromPrimary ? prisma : $replica,
runStore,
});
```
The same flag skips constructing the lag estimator, since probing a
replica the feed no longer reads would be measuring the wrong thing.
Independently, `AuroraReplicaLagSource` detected Aurora by letting
`aurora_replica_status()` fail, on the assumption that the app-level
catch made that free. It isn't: an unresolvable function is a query
error the driver reports to the error log on every sample, so a
non-Aurora replica produced a continuous stream of error events while
the estimator quietly fell through to its next candidate. It now
resolves the function with `to_regproc` and memoizes the answer, so the
unparseable call never reaches the wire.
|
||
|
Chris Arderne
|
6997aeb05e |
fix: security release 2026-07-08 (#4316)
⚒️ Publish Worker (v4) / build (supervisor) (push) Has been cancelled
|
||
|
Daniel Sutton
|
ae96b6c175 |
fix: read-your-writes + global-scope idempotency correctness under the run-ops split (#4284)
## What & why
Two related correctness fixes for the run-ops DB split. Under the split,
run-store reads can route to a **lagging read replica**; a just-written
run/waitpoint/batch can then be missed, causing a wrong decision.
**1. Read-your-writes → owning primary.** Surfaced first as an
intermittent `wait.until({ idempotencyKey })` re-wait on retry. Auditing
the run-store read surface found the same class at sibling sites (some
gating mutations or returning spurious 404s, others
tolerable/self-healing). Reads that must observe their own writes now
route to the owning **primary**
(`findRun`/`findWaitpoint`/`findBatchTaskRunByFriendlyId` →
`*OnPrimary`, a primary re-read on a miss, or a retryable 404 where the
SDK polls). Read-view reads stay on the replica. All additive — the
happy path is unchanged.
**2. Global-scope idempotency across the split.** A `global`-scope key
carries no per-run salt, so the same `(env, task, key)` triggered
concurrently from parents resident on **different** run-ops DBs could
dedup-miss on each DB and create a duplicate (the per-DB unique index
can't enforce cross-DB uniqueness). Such triggers (global scope, or
scope-absent, while split is active) are serialized through the existing
Redis idempotency claim, the loser resolves the winner by id across both
DBs, and the claim is reacquired on the expired/failed
clear-and-recreate path. `run`/`attempt` scope embed the run id and
never contend.
## Stacked for review
This is the **base** of a 2-PR stack, split so review is easier:
- **This PR** — production code only (34 files).
- **Stacked tests PR →
https://github.com/triggerdotdev/trigger.dev/pull/4285** — the
caller-driven guards (55 test files) on top of this branch.
## Validation
Local run-ops split, **both 2-DB and 3-DB**, fresh boot on this branch:
SDK canary 64/71 (only the known concurrency/input-streams/s3 failures),
quarantine sweep **0 unexpected** (340 pass / 16 known / 4 local) in
each topology, dashboard e2e 0 failed. No product regressions.
|
||
|
nicktrn
|
890dd66eb5 |
feat(webapp): route ClickHouse reads to an optional read replica (#4081)
## Summary Adds optional configuration to send ClickHouse read traffic to a separate instance (for example a read replica) while writes stay on the primary `CLICKHOUSE_URL`. This lets operators offload read load (runs list, traces, logs, queries) from the cluster that handles inserts. Fully backwards compatible: with nothing new set, every client resolves to `CLICKHOUSE_URL` exactly as before. ## What it adds - `CLICKHOUSE_READER_URL` (optional): a single reader endpoint that the read-only clients fall back to. Read clients resolve `<own URL> ?? CLICKHOUSE_READER_URL ?? CLICKHOUSE_URL`. The task-events client (which both inserts events and reads traces, spans, and logs) is built as a reader/writer pair so queries use the reader while inserts stay on `CLICKHOUSE_URL`. - `RUNS_LIST_CLICKHOUSE_URL` (optional): a dedicated client for the runs list (dashboard list, runs list API, live reload, child-status counts), so the highest-traffic read path can target its own instance. ## Safety Only read-only clients fall back to the reader: logs, query, admin, runs list, the pending-version lookup, and the realtime run-id resolver. The query page is constrained to read-only (the TSQL parser rejects anything that is not a `SELECT`, and a `readonly` setting is applied). The task-events client routes inserts to the writer and queries to the reader per method, so a write can never reach the reader. Pure-write clients (event inserts, replication) always use `CLICKHOUSE_URL`. Note: this PR targets a baseline branch rather than `main` so the diff stays scoped to the read-replica changes. It will be retargeted to `main` before merge. --------- Co-authored-by: Eric Allam <eallam@icloud.com> |
||
|
nicktrn
|
976171ea16 |
feat(webapp): management API for orgs, projects, members, and settings (#4146)
## Summary
Adds a set of PAT-authenticated management API endpoints so orgs,
projects, members/invites, environment variables, and a few
project/environment settings can be managed programmatically (scripting,
automation) rather than only through the dashboard. Each route is a thin
wrapper over the **existing** service the dashboard already uses, with
the same authorization applied at the route layer - no new business
logic.
## Endpoints
**Organizations**
- `POST /api/v1/orgs` - create an org (`createOrganization`)
- `PATCH /api/v1/orgs/:orgParam` - rename (title)
- `DELETE /api/v1/orgs/:orgParam` - soft-delete
(`DeleteOrganizationService`; keeps the active-subscription guard)
**Members & invites**
- `GET /api/v1/orgs/:orgParam/members` - list members + pending invites
- `DELETE /api/v1/orgs/:orgParam/members/:memberId` - remove a member
(last-member guarded)
- `POST /api/v1/orgs/:orgParam/invites` - invite by email
(`inviteMembers`, sends the invite email)
- `DELETE /api/v1/orgs/:orgParam/invites/:inviteId` - revoke an invite
**Projects**
- `PATCH /api/v1/projects/:projectRef` - rename
(`ProjectSettingsService`)
- `DELETE /api/v1/projects/:projectRef` - soft-delete
(`DeleteProjectService`)
- `PUT /api/v1/projects/:projectRef/default-region` - set the default
region by worker-group name (`SetDefaultRegionService`)
- project GET/list now return `defaultRegion` (worker-group name, or
null when unset)
**Environments**
- `POST /api/v1/projects/:projectRef/:env/pause` and `/resume`
(`PauseEnvironmentService`)
- `POST /api/v1/projects/:projectRef/:env/regenerate-api-key` - rotate
the env secret key (`regenerateApiKey`, RBAC `write:apiKeys`)
- env var create now accepts an optional `isSecret` flag
## Auth & authorization
- All routes authenticate with a **Personal Access Token**
(`Authorization: Bearer tr_pat_...`).
- Org/project routes are built on the PAT route builders in
`apiBuilder.server.ts`: `createLoaderPATApiRoute` (already existed) and
**`createActionPATApiRoute`** (added here - the loader builder had no
mutation counterpart). The builder runs auth, resolves the org/project
role-floor via `context`, and enforces a declarative `authorization`
block using the same RBAC actions the dashboard applies
(`manage:organization` / `read:members` / `manage:members` /
`manage:project`). Handlers keep a membership-scoped query as the floor,
so a non-member gets a 404. This also gives these routes `tenantContext`
user attribution (Sentry) and `ServiceValidationError`-to-status mapping
for free.
- **Membership floor (important).** The OSS RBAC fallback grants a
permissive ability, so `ability.can(...)` can't reject a non-member on
self-hosted. Every handler therefore resolves the target scoped to the
caller's membership (`members: { some: { userId } }`) → 404 for
non-members. `authorization` is the *role* gate; this is the *tenant*
gate. `resolveOrganizationForApiUser`
(`organizationApiAccess.server.ts`) is the org-tier version of the
existing `findProjectByRef` - org-addressed PAT routes are new, so no
such helper existed before.
- Env-tier routes reuse the existing `authorizePatEnvironmentAccess`
(`write:apiKeys`).
### What `createActionPATApiRoute` gives you
A route is pure declaration - the builder handles auth, RBAC,
validation, tracing, and error mapping:
```ts
export const action = createActionPATApiRoute(
{
method: "PUT", // one verb, or ["PATCH", "DELETE"] for multi-verb routes
params: ParamsSchema,
body: SetDefaultRegionRequestBody, // zod-validated
context: async ({ projectRef }) => { // resolve the org for the RBAC role-floor
const project = await prisma.project.findFirst({
where: { externalRef: projectRef, deletedAt: null },
select: { organizationId: true },
});
return project ? { organizationId: project.organizationId } : {};
},
authorization: { action: "manage", resource: () => ({ type: "project" }) },
},
async ({ params, body, authentication, ability }) => {
// auth + authz already enforced. Just do the work.
// `throw new ServiceValidationError("Region not found", 400)` → mapped to that status.
return json({ ok: true });
}
);
```
Handled for you, so handlers stay thin:
- **Method allowlist** - `method` accepts a verb or an array; any other
verb → `405` with an `Allow` header, *before* auth runs:
```ts
const allowedMethods = method ? (Array.isArray(method) ? method :
[method]) : undefined;
if (allowedMethods && !(allowedMethods as
string[]).includes(request.method.toUpperCase())) {
return json({ error: "Method not allowed" }, { status: 405, headers: {
Allow: allowedMethods.join(", ") } });
}
```
- **PAT / user-actor auth** in a single roundtrip → `401` on
missing/invalid/revoked token.
- **RBAC** - `context` computes the caller's role-floor for the target
org/project; `authorization` gates it → `403` with a structured error
body.
- **Sentry attribution** - `tenantContext.enrich({ userId })` so events
from the handler carry the acting user.
- **Typed errors** - a thrown `ServiceValidationError` is mapped to its
`.status` (default 400); anything else → `500`, and expected boundary
errors are logged as `warn` (kept out of Sentry).
- **Validation** - params / query / headers / body are all zod-checked →
`400` with details.
## Notes for reviewers
- Everything wraps an existing service; the intent is API parity for
things that are currently dashboard-only, not new behaviour.
- `createActionPATApiRoute` is new shared infra (the PAT + RBAC mutation
builder that didn't exist). It's self-contained - the loader builder and
existing routes are untouched.
- `@trigger.dev/core` gets one additive field (`defaultRegion` on the
project response, optional/nullable for client-server version skew) -
changeset included, patch.
- `removeTeamMember`'s last-member guard is now atomic (Serializable
transaction via the `$transaction` helper, with retry), so the dashboard
and API both get it server-side. Added a `## Transactions` rule to
`apps/webapp/CLAUDE.md` (always use the `$transaction` helper);
migrating the remaining direct usages is tracked in TRI-11698.
## Open questions
- ~~Is PAT the right auth (vs OAT for automation)?~~ **Resolved: PAT.**
Organization Access Tokens are currently internal-only (used by the
image builder) and not user-accessible, so they can't back this yet.
- Should any of these be gated behind a flag or scope?
- Naming/shape of the routes.
|
||
|
Daniel Sutton
|
a1ca64613b |
fix(webapp): reuse the primary db pool for legacy run-ops when DSNs match (#4253)
## Summary When the run-ops split is enabled, the legacy run-ops database client was always constructed as its own connection pool, even when it points at the same database as the primary (control-plane) client. On setups where those two DSNs resolve to the same physical database, this opened a second, redundant pool and doubled the number of connections used against that database. This change makes the legacy client reuse the primary client's pool whenever their DSNs point at the same database, and only open a separate pool when they genuinely differ. ## Fix A small `sameDatabaseTarget` comparison (host, port, database name, user) decides whether the legacy DSN points at the same database as the primary. When it does, the legacy handle reuses the primary client by reference, so no second pool is opened. When the DSNs diverge, the legacy client is built independently as before, so the split still works once the databases are actually separate. Two smaller changes ride along: - An optional per-pool limit for the run-ops read replica, which connects unpooled and so draws raw backend connections; unset, it falls back to the existing default and behaviour is unchanged. - A startup warning about a missing legacy replica URL is now suppressed when the legacy client shares the primary pool, where it would be misleading. ## Verification Booted the webapp end-to-end in three modes and confirmed the pools opened as expected via the client's own startup logs and live backend connection counts: split off (single pool), split on with a shared database (legacy reuses the primary pool, no doubling), and split on with separate databases (legacy opens its own pool). |
||
|
Chris Arderne
|
64e5d732ad |
chore(webapp,core): remove the unused ResourceMonitor server logging helper (#4244)
The `ResourceMonitor` server-side logging helper is no longer used. It periodically logged the webapp process own memory, disk, and CPU usage behind the `RESOURCE_MONITOR_ENABLED` flag (off by default), and was also exported from `@trigger.dev/core/v3/serverOnly` with no other consumers. This removes the helper, its `@trigger.dev/core` export, the webapp wiring, and the `RESOURCE_MONITOR_ENABLED` env var. The supervisor has its own unrelated `ResourceMonitor` class, which is left untouched. |
||
|
Eric Allam
|
29598a77b8 |
feat(webapp): add option to disable PostgreSQL task-event writes (#4242)
## Summary Adds `EVENT_REPOSITORY_POSTGRES_WRITES_DISABLED` (default off), which makes the task-event store skip all PostgreSQL `TaskEvent` writes. It's for deployments that store task events in ClickHouse (`EVENT_REPOSITORY_DEFAULT_STORE=clickhouse_v2`) and no longer want the PostgreSQL copy. ## How it works The guard sits at the single postgres write boundary, `TaskEventStore.create` / `createMany`, so it covers every write path (OTLP ingestion and run-lifecycle events) with one check. Reads are untouched (`findMany` / trace queries / streaming), so existing PostgreSQL events remain readable. Leave it off unless the default store is `clickhouse_v2`, otherwise task events for any run still routed to PostgreSQL would be dropped. |
||
|
Daniel Sutton
|
bea7e2be90 |
feat(webapp,run-store): route run-graph reads and writes through the run-store router (#4237)
## Summary Run-graph data (runs, batches, waitpoints, and their related tables) can now live in a database separate from the control plane, with every read and write routed to the correct database by each run's residency. This makes reading and writing run data more reliable once the two are split, and is a no-op for single-database installs. ## Design - Run-graph table access goes through the run-store router, which selects the legacy or the new run-ops store per run instead of assuming one shared client. - The legacy run-ops client is now independently pointable, so legacy run data can be served from its own database (and replica) rather than the control-plane connection. - Run-graph writes go straight to the run-graph database instead of being forwarded through the control plane, and replication targets are split so runs in the new database still replicate to analytics without under-counting. - Read-through slots refuse the control-plane client, so a missing residency fails loudly instead of silently reading the wrong database. - Migration `20260710120000_drop_remaining_run_graph_seam_foreign_keys` drops the foreign keys that still crossed the run-graph / control-plane seam, which is what lets the two live in separate databases. The split stays off unless explicitly enabled and the two databases are confirmed physically distinct; startup fails closed otherwise. Verified by running the full dashboard end-to-end suite against both a single-database configuration and a three-database configuration (control plane, the new database, and a physically separate legacy database), with runs on both residencies. No misrouted reads in either configuration. |
||
|
Eric Allam
|
5ba8557a51 |
chore(webapp,core): remove the end-of-life v3 (engine V1) execution stack (#4236)
## Summary v3 (the engine that ran the SDK v3 era, internally `RunEngineVersion.V1`) is end-of-life. Following the removal of the v3 execution apps ([#4194](https://github.com/triggerdotdev/trigger.dev/pull/4194)) and the legacy dev websocket ([#4198](https://github.com/triggerdotdev/trigger.dev/pull/4198)), this removes the remaining v3 execution stack from the server. Clients still on v3 (an old SDK or CLI that has not upgraded) keep getting a clear "upgrade to v4" response. Triggers, batch triggers, reschedules, and deploys that resolve to v3 are rejected with a graceful 4xx pointing at the migration guide, never a 5xx, so a stale client cannot affect server health. Self-hosted instances still running v3 should stay on the 4.5.x release line until they migrate. ## What is removed - The MarQS queue and its shared/dev queue consumers. - The v3 socket.io namespaces (coordinator, provider, shared-queue) and the v3 run lifecycle services (attempt, checkpoint, and batch-resume). - The graphile-worker background job system; all live jobs already run on `@trigger.dev/redis-worker`. - The `DEPRECATE_V3_ENABLED` flag: v3 is now rejected unconditionally, so the flag is gone. - Unused v3 exports from `@trigger.dev/core` (the `v3/zodNamespace` subpath and the legacy socket message catalogs) and the now-dead MarQS environment variables. ## What stays The v4 engine is untouched. The graceful v3 rejection boundary stays, `determineEngineVersion` still detects a v3 project so it can reject it, and the batch service plus batch-completion worker stay for current clients. Live queue concurrency limits and metrics now read from the v4 run engine instead of MarQS, and a brand-new dev environment now defaults to v4. ## Dependency cleanup Removes webapp dependencies left unused by this change: `seedrandom` and `semver` (only the removed v3 code used them) plus a set that was already dead, their orphaned `@types` packages, and two dead files. Adds a `knip:deps` script and a `knip.json` config so unused dependencies can be found the same way going forward. |
||
|
Daniel Sutton
|
c601739d35 |
perf(webapp,run-store): grouped run-ops reads + mint-kind flip grace (#4227)
## Summary Two threads on the run-ops split path. Read path: per-item run reads are batched into grouped queries, a waitpoint's connected-run reads are bounded, and the dedicated-schema relation hydrators fetch only the requested columns instead of whole rows. Retrieve also falls back to the other database when a routed read misses, so a run whose physical residency diverges from its id shape is still found rather than returning a spurious not-found. Fewer and lighter queries on the run read path, with no change to results. Mint-kind flip safety: flipping which database new runs mint to is now a deterministic wall-clock cutover, for both per-org and global flips. For a grace window every process resolves the same database, so a flip cannot route two concurrent triggers that share an idempotency key to different databases (which would bypass the per-database unique constraint and create a duplicate run). Supersedes the earlier #4205 and #4208. Draft: validation in progress. |
||
|
Eric Allam
|
45527e317a |
feat(webapp): opt-in worker pool for OTLP ingest transform (#4232)
## Summary Under high OTLP ingest volume, the whole decode, transform, and enrich pipeline runs on the request event loop, so a single CPU core becomes the ceiling while the rest sit idle. This adds an opt-in worker pool that moves decode, transform, and LLM-cost enrichment onto worker threads, keeping the main thread free for I/O. It is off by default (`OTEL_TRANSFORM_WORKER_POOL_ENABLED`), so behavior is unchanged unless enabled. ## Design Workers do decode, filter, convert, and enrich (including LLM pricing match). The main thread stays the single database reader: it loads the pricing registry and broadcasts the compiled model rows to the workers (re-broadcasting on every reload), so workers never touch the database. The pure transform is extracted into a dependency-light module (no Prisma/Redis/ClickHouse imports) so it can run inside a worker. Importantly, the main thread keeps the existing single consolidated insert path, so ClickHouse insert batching and part count are unchanged. The parallelism buys CPU headroom, not more insert streams (which would add merge pressure). The worker is bundled as a standalone file at build time and ships in the existing image with no Dockerfile change. In local load testing the pool sustained roughly 2.6x the throughput of the single-thread path and kept the main thread responsive under load. |
||
|
Eric Allam
|
5d0e9d9dc5 |
feat(webapp): make the default realtime backend configurable (#4231)
## Summary The default realtime backend was hardcoded to Electric. This adds a `REALTIME_BACKEND_DEFAULT` env var (`electric` | `native` | `shadow`, default `electric`) that chooses the backend for any environment whose org has no `realtimeBackend` override. Behavior is unchanged unless you set it; per-org overrides still win. The default is applied at every point where the per-org flag falls through: the initial value, the flag lookup default, and the error fallback. |
||
|
Matt Aitken
|
b64b54c74e |
feat(webapp): pass database writer and reader config to auth plugins (#4229)
## Summary The RBAC and SSO auth plugins can own their own database client, but they could only read `DATABASE_URL`, so every connection they opened landed on the primary. The host webapp now resolves writer and read-replica URLs from its env (the same fallback chain its own Prisma clients use: control-plane URL first, then the default) and passes them to the plugins at create time via a shared `PluginDatabaseConfig`, along with separate connection limits for writes (default 2) and reads (default 5, tunable via `RBAC_DATABASE_*_CONNECTION_LIMIT` and `SSO_DATABASE_*_CONNECTION_LIMIT`). A plugin can then route hot-path reads (per-request auth checks, login routing) to the read replica and keep only rare mutations on the primary. With no replica configured, or no plugin installed, nothing changes: the OSS fallback ignores the new option and keeps reading through the Prisma clients it is already given. 🤖 Generated with [Claude Code](https://claude.com/claude-code) |
||
|
Eric Allam
|
1a0198cc5e |
perf(webapp,clickhouse): move runs empty-state check to ClickHouse (#4202)
## Summary
The runs page's empty-state check (whether an environment has ever had a
run, which decides between the "getting started" and "no runs match your
filters" states) ran a `findFirst` against the Postgres `TaskRun` table.
This moves it to ClickHouse, the same store the runs list itself reads
from, so the check no longer queries `TaskRun`.
## Design
Only the runs list triggers the check now (via an `includeHasAnyRuns`
flag); the other presenters that reuse `NextRunListPresenter` (API,
schedule detail, waitpoint detail, error group) no longer issue it. When
the list is empty it runs `SELECT 1 FROM task_runs_v2 ... LIMIT 1`
filtered on the full `(organization_id, project_id, environment_id)`
sort-key prefix with a configurable `created_at` lower bound
(`RUN_LIST_HAS_RUNS_LOOKBACK_DAYS`, default 30), so it hits the primary
index and reads minimal granules.
Results are cached in a tiered memory + Redis SWR cache. Only positive
("has runs") results are cached, so an environment with no runs is
always re-checked and its first run shows up immediately.
|
||
|
DKP
|
00ee0751ec |
feat(webapp): proxy PostHog through a same-origin /ph path (#4183)
## Summary posthog-js sent product analytics to PostHog Cloud directly from the browser. This points `api_host` at a same-origin `/ph` path that forwards to PostHog Cloud EU server-side, following PostHog's standard first-party reverse-proxy setup. ## How it works A resource route forwards each request server-side, splitting by path: `/ph/static/*` and `/ph/array/*` go to the asset host, everything else (analytics events, feature flags) goes to the ingest host. It rewrites the `Host` header, strips the `/ph` prefix, and streams the response back. Only PostHog's own cookies are forwarded, so the app session cookie stays first-party. Upstream hosts default to PostHog Cloud EU, overridable via `POSTHOG_INGEST_HOST` / `POSTHOG_ASSETS_HOST`. It also sets `cross_subdomain_cookie` so a single PostHog session is shared across the marketing site and app. Verified locally: static assets return 200 from the EU asset host, and analytics events return 200 through the ingest host. |
||
|
Chris Arderne
|
aa74e68c71 |
feat(sdk): add bulk replay to api and sdk (#4105)
## Summary
Adds SDK and API support for run bulk actions. You can now create bulk
cancel or replay actions from `@trigger.dev/sdk` using run IDs or the
same filters as `runs.list()`, then retrieve, list, poll, or abort the
action by its `bulk_` handle.
Tests, docs, changesets added.
## Design
The dashboard bulk action service now accepts structured filters instead
of reading directly from a dashboard request, so the dashboard and API
share the same creation path. Replay actions created through the API are
attributed with the existing `api` trigger source, while
dashboard-created actions keep `dashboard`.
The SDK exposes the new surface under `runs.bulk.*`, including
`targetRegion` for replay region overrides and cursor pagination for
listing bulk actions.
## Filters and runIds
Nuance on filters. If `filter` is provided, it MUST have at least one
key. This is to remove the footgun of passing no filter and selecting
all runs.
```typescript
{ action: "cancel", runIds: ["run_1"] } // valid
{ action: "cancel", runIds: [] } // invalid, min(1)
{ action: "cancel", filter: { status: "FAILED" } } // valid
{ action: "cancel", filter: {} } // invalid
{ action: "cancel", filter: {}, runIds: ["run_1"] } // invalid
```
|
||
|
Daniel Sutton
|
712c7c3b1a |
feat(webapp): add RUN_REPLICATION_RUN_OPS_DATABASE_URL for the runs-replication source (#4160)
## Summary The run-ops runs-replication source now takes its connection URL from `RUN_REPLICATION_RUN_OPS_DATABASE_URL`, required whenever the run-ops split is enabled. The runs replicator speaks the Postgres streaming replication protocol, which cannot run through a transaction pooler, so it needs its own direct endpoint separate from the app's `RUN_OPS_DATABASE_URL` (which may point at a pooler). When the split is on and this is unset, boot fails via `SplitReplicationMisconfiguredError` rather than silently falling back to a wrong endpoint. |
||
|
Daniel Sutton
|
092b9ef07a |
fix(run-ops): DNS-safe, sortable base32hex run id (replace base62 KSUID) (#4154)
## Problem
The run-ops split mints NEW-store run ids as **27-char base62 KSUIDs**.
The supervisor writes the run id into the Kubernetes pod name
(`runner-<id>`), and pod names must be DNS-1123 labels (lowercase
`[a-z0-9-]`) — so uppercase base62 ids make k8s reject the pod (422) and
**those runs never launch** (they loop in `PENDING_EXECUTING` until the
heartbeat-stall handler nacks them, forever). `.toLowerCase()` can't fix
it: base62 has both `A`(10) and `a`(36) as distinct symbols, so folding
collides distinct ids and destroys sort order.
## Fix: change the encoding, not the structure
Mint a **26-char lowercase base32hex** run id:
```
run_<24-char base32hex core><region char><version char>
[ 6-byte ms timestamp ][ 9 CSPRNG bytes ]
```
- **base32hex** (RFC 4648 §7, alphabet `0-9a-v`): lowercase,
order-preserving, DNS-safe; 15 bytes → exactly 24 chars, no padding.
Hand-rolled encode/decode (no new dependency).
- **48-bit ms timestamp** in the leading bytes → plain string sort ==
creation order at millisecond resolution.
- **72 bits CSPRNG** entropy; PK unique constraint is the backstop (no
retry loop).
- **region / version** are raw positional chars (read via one `charAt`
before decoding/routing), version = `"1"`.
DNS-safe from birth and hyphen-free, so **firekeeper is unchanged** —
`runner-<id>-attempt-N` → strip `runner-`, cut at first hyphen still
recovers the exact id incl. region+version.
## Residency discriminator: length → version char
`classifyKind`/`classifyResidency` (`runOpsResidency.ts`) previously
distinguished NEW vs LEGACY by **id length**. That gets ambiguous with a
third format. It now discriminates on the **version char at a fixed
position** (`isRunOpsIdBody`: 26 chars, `[25] === "1"`, base32hex
alphabet) → NEW; everything else → LEGACY. Total, never throws. The
`Residency` (NEW/LEGACY) contract the routing store consumes is
unchanged; the `"ksuid"` `ResidencyKind` label is retained only because
it's the persisted `runOpsMintKsuid` feature-flag value.
## Scope / verification
- Generator + discriminator in `@trigger.dev/core` isomorphic; mint path
+ all id-shape call sites swept (~40 webapp files); changeset added
(`@trigger.dev/core` patch).
- Core unit tests (encode/decode round-trip + property, generator shape,
ms sort-order incl. intra-second, parse partitioned-vs-legacy,
firekeeper round-trip): **24 pass**. `@trigger.dev/core` builds; webapp
typechecks; format/lint clean.
## Open decisions (flagged, not silently chosen)
1. **Backward-compat**: existing 27-char base62 KSUID runs now classify
LEGACY. On test cloud these are the broken/looping runs that never
completed, so this is acceptable — but worth a conscious call before
prod. No transitional length-recognition added (keeps the discriminator
clean).
2. **Storage collation**: the sort guarantee is byte-order — if the
run-ops id column is `TEXT` with default locale collation it's silently
not honored. Confirm whether `COLLATE "C"` / `BYTEA` is needed on the
run-ops schema.
3. **Region sourcing** wiring — see `regionCharForRegion` /
`REGION_CODES`.
---
## ⚠️ Required migration — deploy in lockstep
This PR renames a persisted feature-flag key/value and an env var. These
are **not** changed by the code alone and must be migrated when this
deploys, or affected orgs silently fall back to `cuid` minting (no crash
— `defaultValue: "cuid"`):
1. **Env var** (terraform): `RUN_OPS_MINT_KSUID_ENABLED` →
`RUN_OPS_MINT_ENABLED` (carry the value over).
2. **DB** `organization.featureFlags`: migrate both the key and value
together:
- key `runOpsMintKsuid` → `runOpsMintKind`
- value `"ksuid"` → `"runOpsId"`
Until an org's flag row is migrated, its `runOpsMintKind` lookup misses
and it mints `cuid` (legacy) — so no NEW-store ids for that org until
the data lands.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|
Daniel Sutton
|
962bc48738 |
feat(run-ops): automatically migrate the dedicated run-ops database (#4150)
## What
Adds the ability to **automatically migrate the dedicated run-ops
database** (the NEW DB in the run-ops split), matching how every other
database in the system is migrated. Follow-up to the run-ops split
activation.
## Changes
- **Migrate runner** — new
`internal-packages/run-ops-database/scripts/migrate.mjs`, exposed as
`db:migrate:deploy` / `db:migrate:status`. Connects via
`RUN_OPS_DATABASE_URL` (the same var the app uses) and expands `${VAR}`
refs like Prisma's dotenv.
- **Self-host** — `docker/scripts/entrypoint.sh` runs the run-ops
migration on boot when the DB is configured, gated by
`SKIP_RUN_OPS_MIGRATIONS`. Single-DB installs never set the URL, so it's
a clean no-op.
- **Single env-var family** — the run-ops DB is now addressed by one
canonical `RUN_OPS_*` family, connect path and migrations resolving the
identical URL:
- `RUN_OPS_DATABASE_URL` (writer) — replaces `TASK_RUN_DATABASE_URL`
- `RUN_OPS_LEGACY_DATABASE_URL` — replaces
`TASK_RUN_LEGACY_DATABASE_URL`
- `RUN_OPS_DATABASE_READ_REPLICA_URL` — replaces
`TASK_RUN_DATABASE_READ_REPLICA_URL`
- the old `TASK_RUN_*` aliases, the `??` coalesce, the
`runOpsNewDatabaseUrl` indirection, and the migrate-only `directUrl` are
all removed (consumers read `env.RUN_OPS_DATABASE_URL` directly).
`directUrl` was dropped because it was only ever used by `prisma
migrate` (never the app runtime) to bypass a pooler for advisory locks —
premature here since the run-ops connection isn't wired to the app yet.
If a pooler is later introduced for the app, a direct URL can be
reintroduced then.
## Safety
- **Pure rename** — nothing deployed sets any `TASK_RUN_*` var (the
split isn't activated anywhere yet; `.env.example`, docker-compose, and
cloud already use `RUN_OPS_*`), so there is no config migration.
- **Single-DB / self-host** — no new required env var; entrypoint and
migrate are no-ops when `RUN_OPS_DATABASE_URL` is unset.
- **Cloud** — runs migrations as pre-deploy ECS tasks (companion cloud
PR), calling these same `db:migrate:deploy` / `db:migrate:status`
commands.
## Verification
- Live migration against a fresh scratch DB with only
`RUN_OPS_DATABASE_URL` set: both migrations applied, no `P1012`/`P1013`;
`${VAR}` expansion, idempotent re-run, `status`, and no-op skip all
pass.
- Schema parity 4/4; `typecheck --filter webapp` 18/18; affected
split/replication tests 34/34.
## Scope
This delivers automatic migrations only. Enabling the app to *use* the
new DB (setting `RUN_OPS_DATABASE_URL` + `RUN_OPS_SPLIT_ENABLED` on the
service) is a separate activation step.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|
Daniel Sutton
|
8465ac5ac3 |
feat(run-ops): webapp db topology, flags, and split-mode resolver wiring (#4117)
## What Wires the run-ops split into the webapp: database topology, environment flags, split-mode gating, and the control-plane resolver/cache layer that the run-store and run-engine seams from the previous PR plug into. - **DB topology & env** (`apps/webapp/app/db.server.ts`, `env.server.ts`, `entry.server.tsx`): adds the run-ops database clients/topology and the environment variables that configure and gate the split. - **runOpsMigration module** (new `apps/webapp/app/v3/runOpsMigration/`): the webapp-side machinery — `splitMode.server.ts`, `controlPlaneResolver.server.ts` + `controlPlaneCache.server.ts`, `readThrough.server.ts`, `crossSeamGuard.server.ts`, `distinctDbSentinel.server.ts`, id-minting helpers (`mintBatchFriendlyId`, `runOpsMintKind`, `resolveInheritedMintKind`), `runOpsCascadeCleanup.server.ts`, the split read gate, and route/unblock catalogs. - **Store/engine wiring** (`app/v3/runStore.server.ts`, `runEngine.server.ts`, `runEngineHandlers.server.ts` + new `runEngineHandlersShared.server.ts`): points the webapp's store/engine construction at the resolver, and factors shared handler logic out so both seams use one path. - **Read-path touch-ups**: `runtimeEnvironment.server.ts`, `eventRepository/index.server.ts`, `taskRunHeartbeatFailed.server.ts`, `engineVersion.server.ts` route their run/environment lookups read-through the resolver. - `413a94511` — interlocks split mode against the native realtime backend so the two aren't enabled in an incompatible combination (see `.server-changes/run-ops-split-realtime-interlock.md`). - `dc74c57fd` — drops the earlier "known-migrated" read layer; residency is determined by id-shape only. ## Why PR5 of the run-ops split stack. This is the webapp foundation layer: it stands up the DB topology, flags, and resolver/cache the rest of the stack depends on, and repoints webapp read paths through the resolver. Additive when the split is not enabled (existing single-DB behavior preserved behind flags); behavior-changing on the read-through paths and the realtime interlock. ## Tests New vitest coverage across `apps/webapp/test/` and colocated `*.server.test.ts` files: db topology, split mode, split read gate, cross-seam guard, mint cutover / flip latency, control-plane cache, control-plane resolver, distinct-db sentinel, read-through loaders (route loaders, run-detail loaders, `findEnvironmentFromRun`), and the run-engine handlers. Testcontainers-backed; no mocks. `pnpm-lock.yaml` synced for the two new webapp deps. ## Notes Draft, **stacked on #4116** (`runops/pr04-store-engine`). Review that first; this diff is against it. Server-change / changeset note to be added at stack-assembly time. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
Eric Allam
|
ea9d6f7430 |
feat(webapp): pin the dashboard agent to a deployed version (#4128)
## Summary The dashboard agent runs as a `chat.agent` in its own Trigger.dev project, deployed separately from the app that starts its sessions. This adds independent, non-disruptive deploys for it: a new workflow deploys the agent with `--skip-promotion` (so a deploy never becomes the current version on its own), and the app pins its sessions to a chosen version. ## How it works `DASHBOARD_AGENT_VERSION` (unset by default) pins agent sessions to a specific deployed version; when unset, sessions run on the project environment's current version. The pin is passed on session start and head start and is forwarded to every continuation run, so a pinned session stays on its version for its whole life. Cutting over to a new build becomes a config change (set the version) rather than a redeploy, and rollback is flipping it back. The workflow (`dashboard-agent-deploy.yml`) runs a leg per environment (staging and prod), each gated by its own environment, and triggers on pushes to `main` that touch the agent or its store (also available via manual dispatch). Deploy versions are per-environment, so each environment pins to its own leg's version. |
||
|
Oskar Otwinowski
|
c7f6ed501c |
feat(cli,core): add opt-in dev-only telnet log streaming (#4110)
Stream dev logs over a local telnet/TCP socket. `trigger dev` mirrors its terminal output on port 6767 by default (override with --telnet-logs-port or TRIGGER_DEV_TELNET_LOGS_PORT, 0 disables). webapp, supervisor, and coordinator each expose an opt-in stream gated on a per-service *_TELNET_LOGS_PORT env var. New @trigger.dev/core/v3/telnetLogServer module (localhost-only, backpressure-safe, plain-text) plus optional static Logger.onLog / SimpleStructuredLogger.onLog sinks. Then you (or your agent) can use `nc` to connect and filter out the stream. <img width="1103" height="239" alt="image" src="https://github.com/user-attachments/assets/b4d47efc-8a57-4185-a159-10f2806627ae" /> |
||
|
Eric Allam
|
f1bd11a7ef |
feat(webapp): gracefully shut down the v3 engine behind a flag (#4017)
## Summary Adds a single env flag, `DEPRECATE_V3_ENABLED` (default off), that gracefully winds down the v3 engine (`RunEngineVersion.V1`). While it's off nothing changes, so self-hosted instances still on v3 keep working. When it's on: - Triggers that resolve to v3 are rejected with a clear, actionable error pointing at the [v4 migration guide](https://trigger.dev/docs/migrating-from-v3), instead of silently creating runs that never execute. This covers single triggers, batches, scheduled fires, replays, and `triggerAndWait`, which all funnel through one place. - The legacy `trigger dev` websocket used by v3 CLIs is closed with an upgrade message (v4 CLIs use a different dev transport). - The v3 shared-queue consumer refuses to start, so no deployed v3 runs are dequeued. - The v3 run-lifecycle background jobs (heartbeat timeout, TTL expiry, retry, resume batch/dependency, delayed-run enqueue, and scheduled fires) become no-ops, so abandoned v3 runs stop generating database load. This builds on the existing deploy deprecation flag, which already rejects v3 CLI deploys. ## Design Enforcement is read through one helper, `isV3Disabled()`. Every gate combines it with a per-run or per-project engine check (`isV3Disabled() && engine === "V1"`), so a v4 run that happens to reach a shared service behaves exactly as before. v4 (V2) is never affected. The flag is a hard switch, not a drain: when it's on, in-flight v3 runs are abandoned in place rather than failed or expired, which is the intended behaviour for the final shutdown. |
||
|
Katia Bulatova
|
b1987dc090 |
feat(webapp): billing limits — pause, reject, recovery, and settings UI (#3996)
## Summary Adds Billing Limits to the webapp. Customers can set a monthly spend cap. When usage crosses the limit, billable environments enter a grace period. If the limit is not resolved before grace expires, new triggers are rejected until the organization increases or removes the limit. |
||
|
Chris Arderne
|
4fde283e76 |
chore: format and lint webapp also (#4056)
#3977 added formatting and linting everywhere else. This extends it to the webapp. |
||
|
Eric Allam
|
9ef5cf0055 |
fix(webapp): stop showing the in-dashboard agent to admins by default (#4050)
## Summary The in-dashboard agent button was rendered for all admins and impersonators regardless of the `hasDashboardAgentAccess` flag, so it appeared even where the agent is disabled (for example, floating over the run inspector controls). It is now gated by the flag for everyone, so it stays hidden until the flag is turned on. ## Rollout Both levers default off, so nothing changes for users until deliberately enabled: - **Per-org:** set `hasDashboardAgentAccess` on an org's feature flags to enable the agent for just that org. - **All admins:** set `DASHBOARD_AGENT_ADMIN_PREVIEW=1` to give admins and impersonators an everywhere-preview, independent of the per-org flag. Previously admins bypassed the flag unconditionally, which is why the button showed up before the agent was ready to ship. |
||
|
nicktrn
|
bc605eedaf |
fix(webapp): verify deployment image exists before finalizing (#4049)
A deployment could be marked deployed and promoted to current without its image ever landing in the registry. Finalize trusted the CLI: the v1 path never pushed or checked, and the v2/v3 path skips its own push when the CLI sends `skipPushToRegistry` - which the local-build path always does. In the happy path the CLI pushes the image itself, so this stayed latent. But any deviation - `--no-push`/`--load`, a push that lands in a different registry, or an old CLI - promoted a version whose image can't be pulled, so every run failed at pull time while the deploy itself reported success. This adds a registry existence check after push and before finalize. If the image isn't there, the deploy fails loudly instead of promoting a version that can't start. The check is ECR-only (a no-op for other registries, so self-hosted setups are unaffected) and uses `BatchGetImage`, which the deploy role already allows. It fails open on an ambiguous registry error so the check can't itself turn into a deploy outage. The image reference is the platform-generated value and the lookup is bound to the configured registry host; the CLI-supplied digest is validated before use. Can be turned off with `DEPLOY_IMAGE_VERIFICATION_ENABLED=0` for setups that push images out of band (e.g. an air-gapped registry the platform can't reach). refs TRI-11243 |
||
|
Eric Allam
|
2fa84ea124 |
feat(webapp): gate worker dequeues by worker queue via env var (#4030)
## Summary Adds a `RUN_ENGINE_DEQUEUE_DISABLED_WORKER_QUEUES` setting that refuses worker dequeue requests for the listed worker queues (or base regions), so their runs stay queued instead of being handed to workers that can't run them. Blocked dequeues are counted via a `run_engine.dequeue.blocked` OTel counter (labeled by `worker_queue` and `region`). |
||
|
Eric Allam
|
8890d7a258 |
feat(run-engine,webapp): always report worker queue length metrics (#4029)
## Summary The `runqueue.workerQueue.length` gauge only reported a worker queue's depth while runs were being dequeued from it. When dequeues stop, the metric goes stale or missing, so a queue that has backed up because nothing is draining it can't be alerted on. This adds a small observer that refreshes the observed set of worker queues from the `WorkerInstanceGroup` records on an interval, so every active worker queue (and its scheduled split variant) keeps reporting its length regardless of dequeue activity. The observer is off by default and enabled per service via `RUN_ENGINE_WORKER_QUEUE_OBSERVER_ENABLED`, reads from the read replica, and skips a configurable set of cloud providers (`RUN_ENGINE_WORKER_QUEUE_OBSERVER_EXCLUDED_CLOUD_PROVIDERS`, default `digitalocean`). When enabled it is the source of truth for the observed set, so the per-dequeue registration is skipped on that instance, and it groups by worker queue so the per-instance duplicates collapse to the true depth. Also removes the unused `GET`/`POST /api/v1/workers` endpoints. Their only consumer was a CLI command group that is no longer registered. ## Verification Verified end to end against a local stack: the gauge reports each worker queue's length with no dequeues happening, excludes the configured providers, includes hidden groups, and the removed endpoints return as if they never existed. Added a run-engine test (`workerQueueObservation.test.ts`). |
||
|
Eric Allam
|
c06005b353 |
feat(webapp,sdk): in-dashboard AI agent (#4018)
## Summary Adds an in-dashboard AI agent: a chat panel, reachable from any environment page, that answers questions about your runs, errors, tasks, and analytics, diagnoses why a run failed, charts your data, reads your connected repo's source, and answers product and how-to questions. It is gated behind the `hasDashboardAgentAccess` feature flag (global or per-org, default off), so this PR ships disabled: the launcher is hidden unless the flag is enabled. ## Design The agent runs as a standalone `chat.agent` Trigger task in its own internal package, with no access to the webapp database, Prisma, or ClickHouse. It reads the user's data over the public API, acting as the user via a short-lived delegated user-actor token minted server-side each turn (never in the browser), building on [#3997](https://github.com/triggerdotdev/trigger.dev/pull/3997). The error and analytics tools use [#4005](https://github.com/triggerdotdev/trigger.dev/pull/4005) and the TRQL query API. The first turn of a new chat streams from a warm webapp route (Head Start) while the durable agent boots in parallel. Structured answers (a run-failure diagnosis card, a live chart) render through a small typed view catalog rather than arbitrary markup. A knowledge lane forwards product and how-to questions to the support assistant. Conversation history lives in a separate Drizzle-backed store on its own Postgres schema, kept as a display read-model so it can never corrupt the agent's model context. The SDK changes add an `apiClient` option to `chat.createStartSessionAction` and `chat.headStart`, and keep the Head Start tool-approval tail intact across a custom `prepareMessages` hook so prompt caching and Head Start compose. |
||
|
Eric Allam
|
c6f0769299 |
fix(webapp): bound logs search memory and fix pagination at scale (#4012)
## Summary The logs search page (behind a feature flag) ran ClickHouse out of memory when browsing back over long time ranges. This keeps it within bounded memory and fixes a pagination bug that could skip or duplicate rows at a page boundary. ## Fix Memory: the list query reads in sort-key order, which opens one read stream per part in the window, and on object storage those per-part read buffers dominate peak memory, so it scaled with the number of parts scanned. Two changes bound it: - The logs ClickHouse client caps the per-part read buffers via new env-tunable settings. The object-storage-only setting is opt-in, so it is never sent to a ClickHouse version that lacks it. - Recent-first window narrowing: rows come back newest first, so the presenter probes the most recent window and only widens toward the full requested range when a page is short. A busy environment fills a page from a few recent parts instead of scanning the whole range; a quiet one still returns every row in a couple of cheap reads. Correctness: the keyset cursor ordered on (triggered_timestamp, trace_id), which is not unique because the spans of a trace share both, so rows at a tie could be skipped or duplicated across pages. The cursor and ORDER BY now include span_id, and the cursor is versioned so stale cursors reset to the first page. Guards: the effective page size is capped, and the existing per-query memory limit lets a pathological wide browse fail with an error instead of taking the node down. ## ClickHouse 26.2 The memory fix relies on lazy materialization deferring the wide attributes column to the output rows, which only holds on 26.x. Cloud already runs 26.2, so this moves the dev stack, testcontainers, and CI to match. The ClickHouse test suite passes on 26.2. 🤖 Generated with [Claude Code](https://claude.com/claude-code) |
||
|
Oskar Otwinowski
|
e98a547e6c | feat(sso): SAML/OIDC single sign-on (#3911) | ||
|
Iss
|
c97d246197 |
feat(webapp): sync new orgs + users to Attio CRM on signup (#3896)
Pushes new organizations and users into the Attio CRM at signup time, for Customer Success (TRI-10431). - Orgs → Attio `workspaces`, users → Attio `users`, keyed on Attio's built-in unique `workspace_id` / `user_id` so writes are idempotent upserts. - Runs on the common Redis worker (not inline), so a slow or unavailable Attio never blocks the signup path; failures retry (3 attempts). - Hooks: user-created (alongside the existing Loops call) and org-created (`createOrganization`). - Gated behind `ATTIO_API_KEY`, no key means the sync is skipped entirely, so OSS / self-hosted installs are unaffected. Only creation is covered here (the record "shell"); spend, runs, plan changes, churn, and role/relationship linking are populated by the scheduled full sync, tracked separately. **Deploy note:** requires an Attio API key set as `ATTIO_API_KEY` in the webapp env, with scopes **Records (read-write)** + **Object Configuration (read)**, the assert/upsert endpoint reads object config to resolve the matching attribute. Without the key the sync no-ops. --------- Co-authored-by: Matt Aitken <matt@mattaitken.com> Co-authored-by: devin-ai-integration[bot] <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
nicktrn
|
7aa871f37b |
feat(webapp): plan-aware compute migration (#3957)
Adds an opt-in mechanism to route a configurable percentage of organizations onto the compute (MicroVM) backing of their region at trigger time, without changing their stored region settings. Routing is gated by three global feature flags - `computeMigrationEnabled`, `computeMigrationFreePercentage`, `computeMigrationPaidPercentage` - plus a per-org `computeMigrationEnabled` override that wins in both directions. A region's compute backing is resolved from a new `WorkerInstanceGroup.region` column: a container group and its MicroVM group share one geo `region`, so the migration swaps the resolved worker queue to the backing group's queue. Orgs are bucketed deterministically by id, so ramping a percentage down keeps a strict subset rather than reshuffling, and a region with no compute backing is never touched. Everything is off by default - behaviour is unchanged unless the flags are set. The flags and the worker-region groups are read on the trigger hot path from in-memory snapshots rather than the database: a small `createReloadingRegistry` helper loads each at startup and refreshes them on an interval, so no per-trigger query is added and a percentage or kill-switch change propagates within the reload interval. A cold replica whose snapshot hasn't loaded yet reads as not-migrated (the container path) and self-corrects on the next load - the same cold-start contract as the datastore / LLM-pricing registries, with a `reloading_registry_loaded` metric so a never-loaded registry is alertable. The same migration decision is consulted at deploy-time template creation so a migrated org gets a compute template built ahead of its first run. This runs in shadow mode (best-effort, never fails the deploy) by default, or - when the `computeMigrationRequireTemplate` flag is on - in required mode, built synchronously at deploy so the first run never builds on-demand and template errors surface at deploy time. So operators keep "which runs ran where" while customers only see geography: the run's actual worker queue is stored raw, and the geo region is stamped separately on `TaskRun.region` (and a new ClickHouse `region` column) at trigger time. Read surfaces - the dashboard, the API, and the Query/Logs page - show the geo region, falling back to the worker queue for runs written before the column existed. Minor follow-ups left out of scope: the percentage flags render as text inputs on the admin flags page (the catalog UI has no numeric control type yet), and `createReloadingRegistry` could later gain pub/sub for sub-second cross-replica propagation if the reload interval proves too slow. |
||
|
Matt Aitken
|
f48c89752c |
perf(webapp): parallelize streaming batch-item ingest (#3777)
## Problem
The item-streaming endpoint of the two-phase batch API (`POST
/api/v3/batches/:batchId/items`) processed streamed items strictly
sequentially. For a batch of many large payloads, each offloaded to
object storage inline, this serialized N object-store round-trips inside
a single request and could exceed Node's default `server.requestTimeout`
(300s). The webapp then returned `408`, which the SDK reads as `408
terminated` and retries up to 5 times, turning a slow ingest into a
failure that takes tens of minutes to surface.
## Fix
Ingest now runs through `p-map` over the NDJSON async iterable with
bounded concurrency (`STREAMING_BATCH_INGEST_CONCURRENCY`, default 10):
- `p-map` pulls lazily from the stream, so at most `concurrency` items
are read and in-flight at once. Peak memory stays bounded to roughly
`concurrency × STREAMING_BATCH_ITEM_MAXIMUM_SIZE` and request-body
backpressure is preserved.
- Set the env to `1` for fully sequential ingestion (escape hatch).
## Why this is safe (ordering and idempotency unchanged)
- Ordering derives from each item's index (enqueue `timestamp =
batch.createdAt + index`), not enqueue order.
- Dedup is atomic per index in `enqueueBatchItem`.
- The NDJSON parser now stamps oversized-item markers with their emit
position, removing the consumer's sequential `lastIndex` assumption (the
only order-dependent bit).
- The count-check and conditional-seal path is untouched.
## Scope
This speeds up every batch ingested through the streaming endpoint, not
just large-payload batches. Each item does a per-item Redis enqueue
regardless of size, and those now overlap. Large payloads benefit most
because they add an object-store offload round-trip on top of the
enqueue.
## Verification
Added an integration test (`streamBatchItems.test.ts`) that drives the
real service against Postgres + Redis + RunEngine and times a 150-item
batch at increasing concurrency. Object-store offload is modelled as a
fixed per-item latency (local round-trips are too small to compare
meaningfully):
```
runCount=150
large payloads (10ms/item offload):
concurrency=1 1739ms
concurrency=10 192ms (9.1x faster)
concurrency=50 57ms (30.7x faster)
small payloads (Redis enqueue only, no offload):
concurrency=1 90ms
concurrency=10 24ms (3.7x faster)
```
The test asserts correctness at every concurrency (all items accepted,
sealed, enqueued exactly once), that parallel ingest beats the
sequential floor, and that the small-payload case is strictly faster
than sequential, so the win is not specific to large payloads.
Also exercised end-to-end over real HTTP against a local server: a
20-item batch (12MB body) ingests and seals, a re-stream of the sealed
batch returns `sealed: true` with zero re-accepted items (idempotent
retry), and an oversized item still seals at its correct index.
Existing coverage stays green: concurrent ingest of a 100-item batch,
in-flight processing never exceeding the configured concurrency,
concurrent dedup on streaming retry, and emit-position marker indexing.
## Follow-ups (not in this PR)
- SDK pre-offload of large item payloads (send `application/store` refs
instead of raw blobs) to remove object-store work from the request hot
path and shrink the request body.
- Optional `server.requestTimeout` bump as a safety net.
## CI fix
Added `.github/workflows/codeql.yml` to replace GitHub's automatic
("dynamic") CodeQL scanning. The dynamic setup was failing to upload
SARIF results because the auto-generated `GITHUB_TOKEN` lacked the
`security-events: write` permission. The explicit workflow grants that
permission at the job level and pins all actions to commit SHAs,
consistent with the repo's security conventions.
## ✅ Checklist
- [ ] I have followed every step in the [contributing
guide](https://github.com/triggerdotdev/trigger.dev/blob/main/CONTRIBUTING.md)
- [ ] The PR title follows the convention.
- [ ] I ran and tested the code works
---
## Testing
- Integration test (`streamBatchItems.test.ts`) validates correctness
and performance at concurrency 1, 10, and 50 for both large and small
payloads.
- End-to-end verified over real HTTP: 20-item/12MB batch ingests and
seals, idempotent retry returns `sealed: true`, oversized item seals at
correct index.
---
## Changelog
Streaming batch ingest now processes items with bounded concurrency
instead of one at a time, so batches of many large payloads ingest far
faster and no longer time out. Concurrency is configurable via
`STREAMING_BATCH_INGEST_CONCURRENCY` (default 10); set it to 1 for fully
sequential ingestion.
---
## Screenshots
_[Screenshots]_
💯
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
|
||
|
Eric Allam
|
954ee5c572 |
fix(webapp): deliver realtime changes with current content when the read replica lags (#3910)
## Summary When the realtime runs feed (the backend behind the `realtimeBackend` feature flag) hydrates a change from a Postgres read replica, the read can race the replica's apply of the very write that triggered it. The delivered row then carries the previous change's content, and an isolated final change (for example a last `metadata.set` before a run goes quiet) is not corrected until the roughly 20 second backstop poll. Measured against a replica with deliberate apply delay, every delivery trailed exactly one change behind and a final change stranded for the full backstop interval. ## Fix Publishers stamp each change record with the committed row's `updatedAt`, taken from writes they already perform, so the stamp costs no extra queries. The router delays its wake hydrate until the replica's measured lag has passed, anchored to that timestamp: a record that has already spent longer than the lag in transit is hydrated immediately, so only the racing leading edge ever waits. After hydrating, a tripwire compares each row against its record's watermark. Still-stale rows are withheld and retried briefly, and each detection feeds the lag estimate. If retries run out, the rows are delivered anyway (liveness over freshness) and follow-up re-hydrates emit the fresh version through the normal working-set diff once the replica catches up, with the backstop as the terminal net. Replica lag is sampled reader-side only, and only while feeds are active. Aurora reports live lag via `aurora_replica_status()`; vanilla Postgres can only report "caught up or not" (mid-apply lag is not honestly measurable from a replica), so tripwire observations floor the estimate there. Deployments without a replica resolve to zero lag and skip the gate entirely. Tunables live under `REALTIME_BACKEND_NATIVE_REPLICA_LAG_*`, and `realtime_native.stale_hydrates` plus `realtime_native.replica_lag_estimate_ms` make replica health observable. Two adjacent fixes: a metadata update that writes nothing no longer publishes a change record, and buffered parent and root metadata operations now publish when the flusher writes them, so those changes wake live feeds instead of waiting for the backstop. For local testing, `docker-compose` gains an opt-in `database-replica` service (compose profile `replica`) with a configurable `recovery_min_apply_delay`, which reproduces replica-lag behavior deterministically. With the gate disabled this rig reproduces the one-change-behind delivery exactly; with it enabled, deliveries arrive with current content at roughly the true replica lag, across write rates faster and slower than the lag itself. |
||
|
Eric Allam
|
187c0476c3 |
perf(webapp): shrink run trace loader payload and add trace span cap controls (#3906)
## Summary The run trace page loader serialized every span's raw OTel events (with full properties) into the response, even though the tree UI only renders the derived `timelineEvents` and the span detail panel refetches what it needs. On event-heavy traces that inflated both the loader payload and the server-side heap copies built per request. This PR keeps raw span events server-side and pairs that with a few related trace-view improvements: - A new optional `TRACE_VIEW_EMERGENCY_SPAN_CAP` env var (unset by default) clamps the trace summary and detailed trace summary span limits on both event store paths, including the public run trace endpoint, so operators can bound trace query sizes in one place without retuning the per-store limits. - The TreeView virtualizer resolved every rendered row with a linear scan over the whole tree (and `getNodeProps` did the same via `findIndex`); rows now resolve through memoized id lookup maps, which matters once traces reach tens of thousands of spans. - The run stream SSE lookup now applies the same organization membership scoping as the rest of the run page presenters, for consistency. Behavior is unchanged by default: the trace tree renders from the same `timelineEvents` it always has, and the new cap only takes effect when set. |
||
|
Eric Allam
|
f9d57d3bd5 |
feat(webapp): add a new backend for the realtime runs feed (#3864)
## Summary Adds a second backend for the realtime runs feed (`useRealtimeRun`, `subscribeToRunsWithTag`, `subscribeToBatch`), built to stay healthy when a single busy environment has many subscribers watching many runs at once. It is gated behind a feature flag with the existing backend as the default, so nothing changes for users until it is enabled per environment. ## Design A run change is published once, as a small self-describing record, to a single per-environment channel. Every feed is then a predicate over that one stream rather than owning a channel: - A per-instance router indexes the currently-held feeds by run, tag, and batch. When a run changes it hydrates the affected rows once and serializes them once, then fans the result to every matching feed. One hot shared tag watched by many subscribers costs a single database query and serialize, not one per subscriber. - Feeds that don't match a change are never woken, wake delivery per environment is coalesced on a leading edge (250ms default) so a burst of changes costs one wake, and cold reads coalesce onto a single short-TTL-cached resolve. - An admission gate bounds how many cold ClickHouse resolves run concurrently, so a mass reconnect across many distinct filters queues instead of stampeding the database. - Changes that land while a client is between long-polls are delivered on its next poll instead of waiting for the periodic backstop: each environment buffers its recent change records, subscriptions linger briefly after the last feed closes, and a newly-armed poll replays exactly the connection's gap. - The per-connection replay cursors behind that are shared across instances via Redis (a single timestamp each), so a poll landing on a different instance behind the load balancer still reads the connection's true gap instead of falling back to a cold resolve. Cursor reads have a bounded deadline and degrade to the cold-read path on any Redis trouble. - Tag subscriptions with multiple tags match runs carrying all of the tags, mirroring the existing backend's filter semantics, and live long-polls hold for about 20 seconds to match its cadence. - The per-environment channel supports Redis Cluster sharded pub/sub, so the wake path scales horizontally across shards by environment. - The backend reports its health through OpenTelemetry metrics (delivery lag, poll resolution paths, backstop outcomes, replay and cursor-store activity), with a provisioned Grafana dashboard for local development. Everything is behind the feature flag and tunable via env vars; the existing backend remains the default. |
||
|
Daniel Sutton
|
6afc9bfa4c |
fix(run-engine): retry getSnapshotsSince on the replica then primary when the read replica lags (#3889)
## Summary
When `RUN_ENGINE_READ_REPLICA_SNAPSHOTS_SINCE_ENABLED` is on,
`RunEngine.getSnapshotsSince` reads from the read replica. During write
spikes the replica can briefly lag, so the snapshot id a runner just
learned from the writer isn't visible there yet: the lookup threw, the
worker route returned a 500, and the runner waited for its next poll —
turning sub-second snapshot notifications into poll-interval latency
exactly when things are busiest. This PR makes the flag safe to enable:
a replica miss of the since snapshot gets one jittered retry on the
replica (most lag windows are shorter than the ~50–200ms wait, so the
writer is never touched), then falls back to the primary, observed via a
new `run_engine.snapshots_since.replica_miss` counter with an `outcome`
attribute (`replica_retry` vs `primary`). Only genuine misses — absent
on the primary too — remain errors.
## Design
- `getExecutionSnapshotsSince` now throws a typed
`ExecutionSnapshotNotFoundError` so the engine can distinguish the
expected lag miss from real failures. The message string is unchanged
and the error never leaves the engine.
- The recovery path only engages when the flag is on, a distinct replica
client is configured, and no transaction client was passed. With the
flag off, the path is behaviorally identical to before.
- Retry delay bounds are configurable
(`RUN_ENGINE_SNAPSHOTS_SINCE_REPLICA_RETRY_MIN_MS`/`MAX_MS`, default
50/200; `MAX_MS=0` skips the replica retry and goes straight to the
primary).
- The warn log fires only when the primary serves the read (the writer
spill is the operationally interesting event); replica-retry recoveries
are counted but quiet. A permanently-missing snapshot id stays an
error-level failure with a `failedDuring` field, so lag metrics aren't
polluted by bogus ids.
- Stale-tail lag (replica has the since snapshot but not newer rows)
deliberately still returns the replica's view; the next poll catches up.
- The since-snapshot anchor lookup is now scoped to the polled run
(`where: { id, runId }`), so a snapshot id from a different run raises
not-found instead of silently anchoring a too-wide window of the run's
snapshots.
## Test plan
All vitest + testcontainers, no mocks. A new `schemaOnlyPrisma` fixture
(migrated-but-empty clone database) simulates a replica that hasn't
caught up, and a real in-memory OTel meter pins the counter semantics
per outcome.
- [x] Replica catches up during the jittered retry window → served by
the replica, `outcome=replica_retry` = 1, primary never consulted
- [x] Replica permanently missing the since snapshot → served by the
primary, `outcome=primary` = 1
- [x] Snapshot missing on both replica and primary → null, counter = 0
- [x] Replica has the since snapshot but lags by one → the replica's
view is served, no fallback (verified discriminating power: the test
fails if reads secretly hit the primary)
- [x] Flag off with a replica configured → primary serves the read
- [x] Transaction client provided → bypasses the replica entirely
- [x] Since snapshot belonging to a different run → null
- [x] Existing getSnapshotsSince + waitpoints suites green; run-engine,
testcontainers, and webapp typechecks pass
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
|
||
|
Eric Allam
|
85886b96da |
feat(webapp,supervisor): isolate scheduled runs on a dedicated worker queue (#3839)
## Summary Scheduled runs and their descendants can now be routed to a dedicated per-region worker queue, processed by a separate worker fleet, so a burst of scheduled crons no longer competes with standard and agent runs for the same queue and inflates their startup latency. It is off by default and enabled per organization via a feature flag (with a global default), so nothing changes until it is turned on. ## Design At trigger time, any run whose lineage originates from a schedule (`rootTriggerSource === "schedule"`, which already propagates from a scheduled run down to all of its children) gets its worker queue suffixed with `:scheduled`. The worker queue name is an opaque string persisted on the run and used verbatim by enqueue and dequeue, so this needs no Lua, message-envelope, or concurrency changes. Concurrency stays keyed by environment and queue, not by worker queue. On the consumer side, the dequeue endpoint gains an optional `queueClass` selector. A supervisor sends `queueClass: "scheduled"` and the server derives the actual queue from the worker's own group, so a token can only ever reach its own region's queues. A fleet picks its class with the `TRIGGER_WORKER_QUEUE_CLASS` env var (`default` or `scheduled`), so a dedicated scheduled fleet can run alongside the standard one. Verified end to end against a local managed-worker setup: scheduled runs route to the dedicated queue, are drained only by the scheduled fleet, and standard runs are left untouched. |
||
|
Daniel Sutton
|
4ea3ef138f |
chore(webapp,redis-worker): make mollifier constants configurable (#3822)
## Summary The mollifier had ~21 behavioural constants baked in as hardcoded values — the buffer's ack-grace TTL and Redis retry/reconnect tuning, the drainer's poll interval and backoff envelope, the pre-gate idempotency claim TTL/wait/poll, the buffered-run mutate-with-fallback wait loop, the metadata CAS retry budget and backoff, the stale-sweep scan bounds, and the draining-gauge interval. None could be adjusted without a code change, which makes tuning the system under production load impossible. This exposes all of them as `TRIGGER_MOLLIFIER_*` environment variables, each defaulting to its previous hardcoded value. Behaviour is identical unless an operator sets a var, so it's a safe no-op deploy. ## Design The package-level classes (`MollifierBuffer`, `MollifierDrainer` in `@trigger.dev/redis-worker`) gain optional constructor options defaulting to the old constants — backward compatible, hence a patch changeset. The webapp factories and worker bootstraps read the env and pass them through. The route- and concern-level pure helpers (mutate-with-fallback, metadata mutation, idempotency claim, stale-sweep state) keep their existing `?? DEFAULT` option fallbacks and are fed env values at their call sites, so they stay unit-testable without importing `env.server`. ## Test plan - [x] `@trigger.dev/redis-worker` builds - [x] webapp typecheck passes - [x] mollifier buffer + drainer testcontainer suites pass (modulo a couple of pre-existing flaky timing tests) - [x] Reviewer: confirm the `TRIGGER_MOLLIFIER_*` env var names match ops conventions --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> |