- Added the `askExpert` tool to delegate complex questions to a specialist agent, improving the assistant's ability to handle in-depth inquiries.
- Updated the AIChatMessages component by removing the feedback bar and streamlining message rendering.
- Enhanced the AIChatToolCall component with a new label for the expert consultation process.
- Introduced a new reasoning agent to facilitate complex reasoning tasks and integrated it into the assistant's toolset.
- Updated the router prompt to reflect the new capabilities of the assistant, emphasizing the use of the `askExpert` tool for complex queries.
- Renamed the getQuerySchema tool to getTableSchema for clarity and updated its implementation to fetch table-specific column schemas.
- Adjusted the dashboard assistant and related tools to utilize getTableSchema in the workflow for executing TRQL queries.
- Introduced a new generated schema file for TRQL tables to support the updated getTableSchema functionality.
- Updated documentation to reflect changes in the toolchain and usage instructions for querying data.
- Introduced the getApiDetails tool to fetch the full schema for REST API operations, including parameters and request body shape.
- Updated the dashboard assistant to incorporate getApiDetails in the multi-step tool chain.
- Enhanced error messages in executeApiCall to suggest using getApiDetails for missing parameters.
- Updated documentation to reflect the new workflow for API operations, emphasizing the importance of calling getApiDetails before callApi.
- Sparkle hover animation on the Ask AI button and assistant header, plus a
redesigned header close button (Button + ExitIcon + esc shortcut)
- AIChatPanel resumes an existing chat and sends the current page context when
starting a session
- Every completed tool call is shown as an agent step in the transcript, in the
order the agent ran them, labelled with the tool's name and expandable to
reveal its input and output
- AIChatProvider guards history switching against out-of-order responses and
sets messages before the chat id so useChat picks them up in one render
- resources.ai-assistant slug validation + error handling; chat history now
uses a typed Prisma query instead of raw SQL
- buildToolContext validates required slugs
- chat persistence uses sessions.start() in the preload hooks with a lazy
aiChat upsert in onTurnStart
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Updated AskAI component to hide when the chat drawer is open.
- Modified AIChatContextBanner for better styling.
- Refactored AIChatHeader to include scroll handling and improved chat history display.
- Enhanced AIChatInput to support auto-growing text area and focus management.
- Improved AIChatMessages to handle scrolling and display messages more effectively.
- Added error handling and retry functionality in AIChatPanel.
- Introduced new suggested prompts with animation effects.
These changes aim to provide a more seamless and user-friendly chat experience in the AI assistant interface.
## Summary
Dashboard surfaces handle buffered runs by falling back to the mollifier
snapshot:
- Run detail, span detail, streams view (`_app.../runs.\$runParam`,
`resources.../spans.\$spanParam`, `resources.../streams.\$streamKey`).
- Redirect routes (`@.runs.\$runParam`, `runs.\$runParam`,
`projects.v3.\$projectRef.runs.\$runParam`).
- Action routes — cancel / replay / idempotency-reset / debug — under
`resources.taskruns/...` and `resources.../idempotencyKey.reset`.
- Logs download.
- Realtime subscription route + per-run resource
(`realtime.v1.runs.\$runId`, `resources.../realtime.v1.*`).
- `CancelRunDialog` gains an `onCancelSubmitted` callback so submit
isn't raced by the Radix `DialogClose` wrapper.
Stacked on the mutations PR.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierRealtimeRunResource.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierRealtimeRunResourceBuffer.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierRealtimeSubscription.test.ts\` passes
- [x] Manual smoke: trigger a buffered run, open it in the dashboard,
replay/cancel from the UI
---
## Ship-gate follow-up fixes
- **Auto-redirect to root span on direct nav** — loader sets `?span=`
from root span (PG) or buffered snapshot spanId before 302'ing, so
bookmark/share-link/direct-nav doesn't leave the panel collapsed.
- **RunPresenter switches from `findFirstOrThrow` to `findFirst` + typed
`RunNotInPgError`** — kills the per-poll `PrismaClient error` log spam
for buffered runs without changing the route-loader's fallback flow.
- **Span detail panel renders for buffered runs** — `SpanPresenter.call`
now falls back to `findRunByIdWithMollifierFallback` +
`buildSyntheticSpanRun` instead of returning undefined and triggering
the "Event not found" toast loop.
- **Logs download for buffered runs returns a gzipped placeholder line**
— replaces the 404 with a content-encoded line explaining the run is
queued. Same org-membership gate as the PG path.
- **Admin Debug-Run button hidden for buffered runs + SpanRun circular
type alias broken** (squashed) — buttons gate on a new `isBuffered` flag
on the synthetic SpanRun. Required grounding SpanRun in
`SpanPresenter.getRun` to break a circular type alias TS no longer
tolerates once `isBuffered` is a literal field on the shape.
- **Replay action requires user auth + org-membership** (🚩 Devin
finding) — `action` was unauthenticated and the PG `findFirst` had no
org filter, so any caller with a valid `runParam` could replay any run.
Buffered fallback inherited the same gap. Fixed to mirror the cancel
route.
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
Cancel, replay, reschedule, metadata, tags, and idempotency-key-reset
now succeed against a run that's still in the mollifier buffer.
Mutations are applied to the buffered snapshot via Lua CAS; the drainer
carries the mutation forward when it replays.
Primitives added:
- `mutateWithFallback` — PG-first / buffer-fallback resolver with
bounded-wait safety net for entries that transition mid-mutation.
- `applyMetadataMutation` — buffered metadata PUT mirroring the PG-side
retry loop with CAS atomicity.
- `resolveRunForMutation` — discriminated-union resolver used by route
`findResource` so the route builder's pre-action 404 check sees buffered
runs.
Routes wired (whole files, no GET/POST splits):
- `api.v2.runs.\$runParam.cancel.ts`
- `api.v1.runs.\$runParam.replay.ts`
- `api.v1.runs.\$runParam.reschedule.ts`
- `api.v1.runs.\$runId.metadata.ts`
- `api.v1.runs.\$runId.tags.ts`
- `resetIdempotencyKey.server.ts`
Stacked on the reads PR.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierMutateWithFallback.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierApplyMetadataMutation.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierResolveRunForMutation.test.ts\` passes
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
Synthesise QUEUED/FAILED responses from the mollifier buffer when a
TaskRun row hasn't landed in Postgres yet. Wires the synthesis into:
- `ApiRetrieveRunPresenter`
- v1 trace GET route
- v1 spans GET route
- attempts route gains a GET loader (fixes pre-existing Remix "no
loader" 400)
The `readFallback` infra itself lives on the trigger PR (consumed by
`IdempotencyKeyConcern`); this PR adds the route-level
synthetic-rendering primitives.
Stacked on the replay PR.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierSyntheticRedirectInfo.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierSyntheticSpanRun.test.ts\` passes
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
The replay side of the mollifier:
- `DrainerHandler`: reads buffered snapshots and replays them through
`engine.trigger` to materialise PG rows.
- `RunEngine.createCancelledRun`: new public method the handler uses to
write CANCELED rows directly from snapshots (bypass queue + waitpoint,
emit `runCancelled`). Tolerates the cjson empty-table tags edge case
found during validation.
- Drainer fairness: org → env rotation so a heavy env doesn't starve
light ones in the same org.
- Stale-entry sweep + telemetry + alertable gauge so a stuck/offline
drainer surfaces in alerts.
Both the drainer and sweep default-off; nothing fires unless flagged on
(`TRIGGER_MOLLIFIER_DRAINER_ENABLED`,
`TRIGGER_MOLLIFIER_STALE_SWEEP_ENABLED`).
Stacked on the trigger-time decisions PR.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierDrainerHandler.test.ts\` passes
- [x] \`pnpm run test --filter webapp test/mollifierStaleSweep.test.ts\`
passes
- [x] \`pnpm run test --filter @internal/run-engine
src/engine/tests/createCancelledRun.test.ts\` passes
- [x] \`pnpm run test --filter @trigger.dev/redis-worker
packages/redis-worker/src/mollifier/drainer.test.ts\` passes
---
## Ship-gate follow-up fix
**Drainer writes SYSTEM_FAILURE on max-attempts exhaustion.** Adds an
`onTerminalFailure` callback on `MollifierDrainerOptions` so the
customer's run lands a SYSTEM_FAILURE PG row even when the drainer
exhausts `MAX_ATTEMPTS` on a retryable PG error (previously
`buffer.fail()` was called with no row written → silent data loss). The
callback runs before `buffer.fail()` on every terminal path
(non-retryable AND max-attempts-exhausted), and re-throwing a retryable
error from the callback causes the drainer to requeue rather than fail.
Bumps `@trigger.dev/redis-worker` to a **minor** changeset (additive
option + new exported types). Includes 5 unit tests covering both
terminal causes plus the requeue-on-retryable-callback-failure path and
no-callback back-compat.
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
`concurrencyKey` validation accepted only `z.string().optional()` on the
single-trigger and V2/V3 batch endpoints, and the Phase-2 streaming
NDJSON endpoint accepted `z.record(z.unknown()).optional()` for the
entire `options` field. Callers passing `concurrencyKey: someNumericId`
(e.g. `payload.userId`) either failed schema validation on the first two
paths or sailed through on Phase-2 and then failed downstream at
`prisma.taskRun.create` with `Argument concurrencyKey: Expected String
or Null, provided Int`.
The schema now accepts `string | number` for `concurrencyKey` and
stringifies on the way in, across all three paths. The Phase-2 NDJSON
`options` is tightened to reuse the strict
`BatchTriggerTaskItem.options` shape so it validates identically to the
V2/V3 batch endpoints.
A defensive `typeof === "number"` coercion at the `engine.trigger` call
site in `RunEngineTriggerTaskService` covers in-flight Redis-stored
batch items enqueued before the schema fix — those items are rebuilt
from a `Record<string, unknown>` shape that bypasses the new schema and
would otherwise continue failing for up to their TTL.
## Test plan
- [x] `packages/core/src/v3/schemas/batchItemNDJSON.test.ts` — unit
tests covering numeric→string coercion, string passthrough, no-options,
and rejection of non-string/non-number shapes across
`TriggerTaskRequestBody`, `BatchTriggerTaskItem`, and `BatchItemNDJSON`.
- [x] `apps/webapp/test/engine/triggerTask.test.ts` — `containerTest`
simulating the in-flight Redis batch-item shape (numeric
`concurrencyKey` via `Record<string, unknown>`), verifies the run is
created with `concurrencyKey: "51262"`. Without the worker coercion, the
test reproduces the production stack at `prisma.taskRun.create`.
- [x] `pnpm run typecheck --filter webapp` clean.
- [x] `pnpm run build --filter @trigger.dev/core --filter
@trigger.dev/sdk --filter trigger.dev` clean.
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
The trigger hot path's mollifier integration:
- `mollifyTrigger`: when the gate trips, write the engine.trigger
snapshot to the buffer and return a synthesised QUEUED response.
Postgres write is deferred to drainer-replay (next PR in the stack).
- Pre-gate idempotency-key claim: same-key triggers serialise through
Redis so a burst lands in PG / buffer exactly once.
- Read-fallback extensions: `findRunByIdWithMollifierFallback` for the
trigger-time idempotency lookup that must see buffered runs.
- Gate bypasses: `debounce`, `oneTimeUseToken`,
`parentTaskRunId`/`triggerAndWait` skip the mollify path entirely.
- `triggerTask` + `IdempotencyKeyConcern` wired to the above.
All behaviour gated by the master `TRIGGER_MOLLIFIER_ENABLED` switch;
off-state hot path is unchanged (the gate is not even consulted).
Stacked on the buffer extensions PR.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter webapp test/mollifierMollify.test.ts\`
passes
- [x] \`pnpm run test --filter webapp
test/mollifierIdempotencyClaim.test.ts\` passes
- [x] \`pnpm run test --filter webapp
test/mollifierReadFallback.test.ts\` passes
- [x] \`pnpm run test --filter webapp test/mollifierGate.test.ts\`
passes
- [x] \`pnpm run test --filter webapp test/engine/triggerTask.test.ts\`
passes
---
## Ship-gate follow-up fixes
- **Batch items bypass the mollifier gate** — fixes
`BatchTaskRunItem_taskRunId_fkey` FK violation on batch triggers when
the gate trips. End-state is a drainer-side `BatchTaskRunItem`
create-on-materialise; batch traffic passes through the gate until that
lands.
- **IdempotencyKeyConcern honours buffered-run TTL on expiry** —
buffered path now clears expired idempotency claims (read-side) and
resets the buffer's `mollifier:idempotency:*` SETNX binding (write-side)
so a re-trigger past the customer's TTL lands as a fresh run instead of
echoing the stale buffered runId.
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
Buffer-side data layer used by the rest of the mollifier phase-3 stack.
- `buffer.ts` gains entry inspection (`getEntry`), idempotency lookup
(`lookupIdempotency`), in-place snapshot mutation (`mutateSnapshot`),
and dwell tracking. All atomic via Lua.
- `mollifierSnapshot.server.ts`: shared `MollifierSnapshot` type plus
(de)serialise helpers.
- Drops the entry-TTL config and its env var. The drainer is the
recovery mechanism; an entry that survives the drainer should surface as
a stale-sweep alert, not silently TTL away.
Adds methods to the buffer interface; nothing consumes them yet.
Subsequent PRs in the stack wire trigger-time mollify, read-fallback,
and mutation paths against this surface.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\` passes
- [x] \`pnpm run test --filter @trigger.dev/redis-worker
packages/redis-worker/src/mollifier/buffer.test.ts\` passes
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Make `POST /api/v1/deployments/:deploymentId/background-workers`
idempotent so client-side retries no longer collide on the
`BackgroundWorker` `(project, env, version)` unique index. Helps make
deployments more resilient against the class of indexing failures that
surfaces in the dashboard as "Indexing timed out", e.g. during transient
database issues.
## Summary
Second class of poisoned-row failure in the runs replication path. PR
#3708 plugged lone UTF-16 surrogates; this one handles bare JSON integer
literals outside ClickHouse's `Int64`..`UInt64` range. Recovery stays
purely reactive — the existing `sanitizeRows` walker just gains an extra
branch, so the hot replication path pays nothing on healthy rows.
Fixes the still-firing customer-facing symptom from
[TRI-9755](https://linear.app/triggerdotdev/issue/TRI-9755):
`scan-social-profiles` runs continued to be stranded in `EXECUTING` on
the Tasks page after #3708 deployed. CloudWatch showed `Dropped batch —
ClickHouse JSON parse error but sanitizer found nothing to fix` firing
**8/8 times** since the previous deploy (zero successful sanitizations).
Root cause: upstream JS Number precision loss on a 21-digit Google Plus
ID (`117039831458782873093` → `117039831458782870000`) — the
precision-lossy value still serialises as a bare integer that exceeds
`UInt64.MAX`, which ClickHouse rejects with `INCORRECT_DATA`.
## How the bug ships
The customer task emits an output containing a Poshmark profile's
`spec_format`:
```json
{"key":"gp_id","proper_key":"Gp Id","value":117039831458782870000,"type":"int"}
```
That value is `1.17e20` — comfortably above `UInt64.MAX` (`1.84e19`) but
comfortably below `1e21`. `Number.prototype.toString` only switches to
exponential form at `|value| >= 1e21`, so `JSON.stringify` emits the
bare token `117039831458782870000` and the ClickHouse
`JSON(max_dynamic_paths)` column fails with:
```
Code: 117. DB::Exception: Cannot parse JSON object here: {…}: (while reading the value of key output): (at row 1)
: While executing ParallelParsingBlockInputFormat. (INCORRECT_DATA) (version 25.12.x)
```
Same error verbatim as prod. The same number quoted
(`"117039831458782870000"`) inserts fine — ClickHouse's dynamic JSON
column accepts a `String` subtype on the same path.
## What changed
`apps/webapp/app/v3/eventRepository/sanitizeRowsOnParseError.server.ts`:
- New private `isUnsafeJsonInteger(value)` helper — true iff `value` is
a finite integer-valued JS Number where `|value| < 1e21` (so
`JSON.stringify` emits integer form, not exponent) **and** `value` falls
outside `[Int64.MIN, UInt64.MAX]`.
- `sanitizeUnknownInPlace` gains a number-branch: when the predicate
holds, replace the Number with `String(value)`. The downstream JSON
column dynamic-types the path as String for that row — fine, since the
value was already precision-lossy upstream (no JS Number above 2^53 is
numerically meaningful anyway).
- Float-valued numbers, large floats (>= 1e21), NaN and Infinity are
left alone — `JSON.stringify` emits them with exponents or as `null`,
both of which ClickHouse accepts.
`apps/webapp/test/sanitizeRowsOnParseError.test.ts`: four new unit tests
+ an extension to `sanitizeRows` covering surrogate + integer fixes
counted together across rows. The unit suite now covers:
- Positive value above `UInt64.MAX` (`117039831458782870000` — the
actual prod value)
- Negative value below `Int64.MIN`
- Boundary values pass through (`42`, `Number.MAX_SAFE_INTEGER`, `2^63`)
- Non-integer numbers untouched (floats, `1e25`, NaN, Infinity)
- The actual `scan-social-profiles` nested shape — finds the offending
`gp_id` deep inside
`output.data.profiles[].spec_format[].platform_variables[].value`
`.server-changes/runs-replication-bigint-recovery.md` — release notes
entry.
## Why reactive, not pre-flight
`#prepareJson` runs millions of times per day on the replication hot
path. Walking every JSON tree to look for oversized integers would add
bounded-but-real CPU on every healthy row. `sanitizeRows` only fires
after a ClickHouse parse-error rejection, which is a few times a day
platform-wide. Extending it costs effectively zero on healthy traffic
and gains us recovery on the rare poisoned row.
## Verification
- Reproduced 1:1 in a throwaway Docker
`clickhouse/clickhouse-server:25.12.11.4` (closest available to the prod
`25.12.1.1579` build). Pre-sanitize JSON fails with the exact prod
error; post-sanitize JSON inserts cleanly and the row is readable with
`gp_id` stored as a String subtype.
- `pnpm --filter webapp exec vitest run
test/sanitizeRowsOnParseError.test.ts` — 22/22 passing (18 existing + 4
new).
- `pnpm run typecheck --filter webapp` — clean.
## Test plan
- [x] `pnpm run typecheck --filter webapp`
- [x] Unit tests pass against new + existing cases
- [x] End-to-end Docker ClickHouse repro confirms recovery
- [ ] Post-deploy: confirm `Sanitizing batch after ClickHouse JSON parse
error` warns fire instead of `Dropped batch …` errors when
`scan-social-profiles` outputs trip CH again
- [ ] Post-deploy: confirm `permanentlyDroppedBatches` counter stops
climbing in
`/stp/trigger-app-prod/ecs/replication/service-container/process-logs`
## What this does NOT do
- Doesn't backfill the ~120k+ existing stranded `EXECUTING` rows in
production. Same as #3708 — that needs a reconciliation/backfill sweep
(separate ticket — TRI-9755 fix#3).
- Doesn't address the upstream root cause (the customer task emitting a
JS-Number-precision-lossy big int). That's a customer-task concern; our
replication path needs to be robust to whatever shape arrives.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
## Summary
Stops the legacy V1 `WAITING_FOR_DEPLOY` drain from running on every V2
deployment promotion. The drain queries `TaskRun` by
`status='WAITING_FOR_DEPLOY'`, which only V1-engine runs ever have — V2
runs use `PENDING_VERSION` and are handled out of band. Calling the
drain on V2 promotes produced empty queries against the status index and
unnecessary reader-DB load.
## Fix
Two layers:
1. Gate the enqueue at the call site in `ChangeCurrentDeploymentService`
so it only fires when the deployment's worker is on engine V1.
2. Add a `LEGACY_RUN_ENGINE_WAITING_FOR_DEPLOY_DISABLED` env var
(default `0`). When set to `1`, the service returns immediately from
`call()` — neuters any jobs already sitting in the worker queue from
before the deploy lands.
V1 customers see no change; V2 promotes no longer trigger the drain.
Linear: [TRI-9864](https://linear.app/triggerdotdev/issue/TRI-9864)
(Urgent)
Production incident:
[TRI-9863](https://linear.app/triggerdotdev/issue/TRI-9863) (mitigated
by image revert in cloud#910)
## Bug
`apps/webapp/package.json` declares `"sideEffects": false`. PR #3333
(`71d98b4e`) replaced the previous real method-call retention idiom at
the two `sessionsReplicationInstance` import sites with:
```ts
import { sessionsReplicationInstance } from "...";
void sessionsReplicationInstance;
```
esbuild treats `void <identifier>;` as a pure expression statement under
`sideEffects: false` and **tree-shakes the entire import** — including
the `singleton(...)` call inside `sessionsReplicationInstance.server.ts`
which is the only thing that fires
`initializeSessionsReplicationInstance()`. The sessions→ClickHouse
logical replication worker never starts, the slot is unconsumed, lag
grows.
### How it manifested in production
cloud#907's image bump rolled the `SessionReplicationService` ECS task
on prod at 14:32 UTC. The new container's startup log emitted `🗃️ Runs
replication service enabled` but **not** `🗃️ Sessions replication
service enabled` or `🗃️ Sessions replication service started`.
CloudWatch `OldestReplicationSlotLag` grew at ~220 MB/min and the `High
replication lag` alarm fired at 14:37 UTC. Prod was reverted to the
previous image (cloud#910) to stop the bleed.
### Verification
`grep` of the built bundle `apps/webapp/build/index.js` (built from
`c0365d36`):
- **3** occurrences of `Runs replication` / `runsReplicationInstance`
strings ✅
- **0** occurrences of `Sessions replication` /
`sessionsReplicationInstance` / `SessionsReplicationService` ❌
The runs path survives tree-shaking because `adminWorker.server.ts` and
`admin.api.v1.runs-replication.*` routes have real method calls
(`.start()`, `.teardown()`, `.backfill()`) — observable uses the
tree-shaker must preserve. The sessions singleton has no real callers,
only the `void` no-ops, hence its complete elimination from the bundle.
## Fix
Replace `void sessionsReplicationInstance;` with an assignment to
`globalThis`, an unambiguous observable side effect the bundler cannot
eliminate:
```ts
(globalThis as Record<string, unknown>).__sessionsReplicationInstance =
sessionsReplicationInstance;
```
Applied at both call sites: `apps/webapp/app/entry.server.tsx` and
`apps/webapp/app/v3/services/adminWorker.server.ts`.
Surrounding comments updated to document the bundler interaction so the
next maintainer doesn't reintroduce `void`.
## Out of scope (follow-ups)
- **Robustness improvement**: change `apps/webapp/package.json` from
`"sideEffects": false` to an allowlist that includes
`*Instance.server.ts` files. Prevents the same regression shape via any
future `*Instance` singleton.
- **Build-time check**: add a `grep` post-build step in `publish.yml`
requiring `"Sessions replication"` to appear in
`apps/webapp/build/index.js`. Catches this exact regression at CI time.
## Test plan
- [x] `pnpm run typecheck --filter webapp` clean
- [ ] After merge + publish: confirm new image's
`SessionReplicationService` container logs `🗃️ Sessions replication
service enabled` and `🗃️ Sessions replication service started` at
startup
- [ ] After re-deploying to prod: confirm `OldestReplicationSlotLag`
stops growing and drains
## Summary
On a ClickHouse `Cannot parse JSON object` rejection,
`RunsReplicationService` now sanitizes lone UTF-16 surrogates across the
failing batch via the existing `sanitizeRows` helper and retries once.
If the sanitizer found nothing or the retry also fails, the batch is
dropped loudly with a counter increment, so the surrounding
`#insertWithRetry` layer doesn't spin three more times on a
deterministic failure. Non-parse errors propagate unchanged.
Mirrors the pattern from #3659 (for `ClickhouseEventRepository`) — same
root cause (lone UTF-16 surrogates in user-provided JSON), same recovery
shape, **reusing the same shared helpers** (`sanitizeRows`,
`isClickHouseJsonParseError`, `parseRowNumberFromError`).
Fixes the customer-facing symptom from
[TRI-9755](https://linear.app/triggerdotdev/issue/TRI-9755): a single
row's poisoned `output` JSON used to take down the
`COMPLETED_SUCCESSFULLY` UPDATE events for its 50+ batch-mates,
stranding them in `EXECUTING` in ClickHouse forever and inflating
"Running" counts on the Tasks page. Confirmed in production this is
ongoing — ~120k stale rows accumulated in a single 5-hour burst on
2026-05-18; smaller continuous leak before and after.
## What changed
`apps/webapp/app/services/runsReplicationService.server.ts`:
- Imports the three helpers from
`~/v3/eventRepository/sanitizeRowsOnParseError.server` (no duplication;
no move).
- New private `#insertWithJsonParseRecovery<T>(rows, doInsert,
contextLabel, attempt)` — generic over `TaskRunInsertArray[]` and
`PayloadInsertArray[]`, structurally identical to
`ClickhouseEventRepository.#insertWithJsonParseRecovery`. Try → on parse
error sanitize the whole batch (the `at row N` hint is logged but not
used to slice — semantics under `input_format_parallel_parsing` aren't
stable) → retry once → drop with loud log if sanitizer found nothing OR
retry still fails.
- `#insertTaskRunInserts` and `#insertPayloadInserts` extract a
`doInsert` closure and hand it to the wrapper. Existing error logging,
span recording, and `recordSpanError` are preserved inside the closure.
- New `private _permanentlyDroppedBatches = 0` counter with a public
getter, for ops dashboards and tests (matches the events-repo
convention). One shared counter for both insert sites — granularity
comes from the `contextLabel` (`task_runs_v2` /
`raw_task_runs_payload_v1`) on every log line.
`.server-changes/runs-replication-utf16-recovery.md` — release notes
entry.
## Why no new tests
The shared helpers already have full unit + real-ClickHouse contract
coverage from #3659
(`apps/webapp/test/sanitizeRowsOnParseError.test.ts`,
`apps/webapp/test/otlpUtf16Sanitization.integration.test.ts`). The new
wrapper is a line-for-line structural port. Adding a parallel
integration test would require synthesizing bad data that *escapes* the
preemptive `detectBadJsonStrings` check in `#prepareJson` but still
trips ClickHouse — non-trivial without hand-crafted fixtures and
wouldn't cover any new logic.
## What this does NOT do
- Doesn't touch the ~120k existing stale `EXECUTING` rows in production.
That needs a reconciliation/backfill sweep (separate ticket — TRI-9755
fix#3).
- Doesn't sanitize the `error` column path
(`runsReplicationService.server.ts:932 const errorData = { data:
run.error };`). Reactive recovery will catch it if it ever poisons a
batch, but feeding it through `#prepareJson` like `output` is a cheap
follow-up.
## Test plan
- [x] `pnpm run typecheck --filter webapp` — clean
- [ ] Post-deploy: confirm `permanentlyDroppedBatches` counter stays at
zero (or near-zero) in
`/stp/trigger-app-prod/ecs/replication/service-container/process-logs`,
and watch for `Sanitizing batch after ClickHouse JSON parse error` warns
to confirm recovery is firing on real traffic
- [ ] Post-deploy: confirm the rate of new
"EXECUTING-but-actually-COMPLETED" zombies in ClickHouse flattens
(current rate ≈ tens-to-hundreds per hour platform-wide)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
## Summary
Two improvements to session `.in/append`:
- Oversize-body 413 responses now carry CORS headers, so browser fetches
see a readable status instead of an opaque `TypeError: Failed to fetch`.
App-side retry-on-disconnect loops no longer spin forever on a
permanently-rejected payload.
- The per-record cap is now computed precisely against S2's actual
ceiling instead of the conservative 512 KiB floor. Legitimate ~600-900
KiB tool outputs (search results, file content) now succeed;
pathological all-quote content that would double under JSON escape still
rejects cleanly.
## Design
S2 enforces a per-record metered size of `8 + 2*H + Σ(header name +
value) + body ≤ 1048576` bytes. With no record headers (our case), the
budget reduces to `body ≤ 1048568`. Verified empirically against cloud
S2 — append succeeds at metered=1048576 and 422s at 1048577 with `record
must have metered size less than 1 MiB`.
The old `MAX_APPEND_BODY_BYTES = 512 KiB` was derived by assuming
worst-case JSON escape doubling (every byte becomes `\"` or `\\`),
giving `(1 MiB - overhead) / 2`. Safe, but rejects ~half the legitimate
input space.
The new flow:
1. Pre-cap the HTTP body at 1 MiB (DoS guard against reading arbitrary
garbage before we can compute the wrap).
2. After reading, `S2RealtimeStreams.#appendPartByName` computes
`Buffer.byteLength(JSON.stringify({data: part, id: partId}), "utf8") +
8` and throws `S2RecordTooLargeError` (a `ServiceValidationError` with
status 413) if it would exceed S2's ceiling. The route's existing error
branch maps the throw to a 413 with a descriptive message.
The 413 CORS fix is a single-line change in `apiBuilder.server.ts` —
`wrapResponse` was being skipped on the body-too-large branch; every
other error branch wraps; the 413 was the exception.
## Test plan
- Empirically verified against cloud S2 with a boundary scan across
`[1048568, 1048569, ..., 1048576]` and across H ∈ {0, 1×5 hdr bytes,
1×14 hdr bytes} — the formula matches exactly
- Browser-side fetch on a 700 KiB POST now resolves with a readable
`status: 413` (no `TypeError: Failed to fetch`)
- A 900 KiB ASCII tool output now passes (would have 413'd at 512 KiB
pre-fix)
## Bug
The `OrganizationDataStoresRegistry` singleton in
`apps/webapp/app/services/dataStores/organizationDataStoresRegistryInstance.server.ts`
was constructed with `$replica`. That client was then used by both the
polling read path *and* by `addDataStore` / `updateDataStore` /
`deleteDataStore` (and their backing `SecretStore.setSecret` upserts).
The write methods route through the read replica, which Postgres rejects
with **error code 25006**:
```
Invalid prisma.secretStore.upsert() invocation:
ConnectorError(ConnectorError { user_facing_error: None, kind: QueryError(PostgresError {
code: "25006",
message: "cannot execute INSERT in a read-only transaction",
...
}), transient: false })
```
User-visible symptom: the admin `/admin/data-stores` "Add data store"
form returns a 400 with this error wrapped, so no
`OrganizationDataStore` row can ever be created via the UI.
The read path (`loadFromDatabase` polling + `SecretStore.getSecret`) is
unaffected because `findMany` + secret read are read-only.
## Fix
Change the registry constructor to take both a writer and a replica:
```ts
constructor(writer: PrismaClient, replica: PrismaClient | PrismaReplicaClient)
```
- `loadFromDatabase()` keeps using `_replica` (and its
`SecretStore.getSecret` calls) — these are background cache-fillers, not
on user-latency-sensitive paths.
- `addDataStore` / `updateDataStore` / `deleteDataStore` (and their
`SecretStore.setSecret` / `deleteSecret` calls) now use `_writer`.
`organizationDataStoresRegistryInstance.server.ts` passes `(prisma,
$replica)` from `~/db.server`. Test sites that constructed with
`(prisma)` now pass `(prisma, prisma)` — the testcontainer exposes a
single client, so the writer/replica split collapses to one connection.
## Files
-
`apps/webapp/app/services/dataStores/organizationDataStoresRegistry.server.ts`
— constructor + read/write split
-
`apps/webapp/app/services/dataStores/organizationDataStoresRegistryInstance.server.ts`
— pass `prisma` alongside `$replica`
- `apps/webapp/test/organizationDataStoresRegistry.test.ts` — 14 call
sites bumped
- `apps/webapp/test/clickhouseFactory.test.ts` — 5 call sites bumped
## Test plan
- [x] Existing `organizationDataStoresRegistry.test.ts` +
`clickhouseFactory.test.ts` still pass (constructor sites updated;
behavior unchanged for tests).
- [ ] After deploy to test cloud, retry `/admin/data-stores` "Add data
store" form for the HIPAA org — should now succeed and the row should
appear.
- [ ] Verify the registry's polling reload picks up the new row within
`ORGANIZATION_DATA_STORES_RELOAD_INTERVAL_MS` (60s default) and the
factory starts routing to the org-scoped instance.
## Summary
When a background worker registers, the engine resolves runs that were
queued before the worker was ready (status `PENDING_VERSION`). That
lookup used to scan a Postgres status index on `TaskRun`. Move it to
ClickHouse: query candidate run ids from `task_runs_v2`, then refetch
the actual rows from Postgres by primary key with a `status =
'PENDING_VERSION'` guard for idempotency.
## Design
The lookup is a pluggable interface on the run engine
(`PendingVersionRunIdLookup`). The webapp wires a ClickHouse-backed
implementation through the org-scoped `clickhouseFactory` using a new
`"engine"` client type, configured by `RUN_ENGINE_CLICKHOUSE_*` env
vars. The URL falls back to `CLICKHOUSE_URL` when unset, so self-hosted
deployments don't need new config to keep working.
When the lookup returns no candidates, one bounded retry is scheduled
~5s later to cover ClickHouse replication lag against `task_runs_v2`.
The Postgres status guard on both the candidate refetch and the inner
`updateMany` prevents double-promotion when a retry races with a
concurrent deploy.
Tests cover three existing PENDING_VERSION cases via a small
Postgres-backed test adapter; new ClickHouse-backed integration tests
will follow.
## Summary
Stamp every Sentry event with the signed-in user and the tenant (org /
project / env) the request belongs to, so "Users Impacted" counts
distinct humans and events become filterable per tenant.
**Design after review (current):**
- `user.id = real user cuid` (from `requireUser`). "Users Impacted"
counts humans, not tenants.
- Tenant context (org / project / env slugs, IDs, env type) moves
entirely onto tags: `org_slug`, `project_slug`, `env_slug`, `org_id`,
`project_id`, `project_ref`, `environment_id`, `env_type`, plus
`impersonating` when set.
- Backed by an `AsyncLocalStorage` scope established at the HTTP entry.
Each entry point fills what it knows; loaders enrich the same scope with
what they already have.
**Zero new database queries.** The middleware does a regex match only.
Dashboard loaders that already query Prisma gain a couple of extra
selected columns; nothing new round-trips.
## How it's wired
- **Express middleware (`tenantContextResolver.server.ts`)** — parses
the URL with a regex and always opens an ALS scope. Populates whatever
subset of slugs is present: `/orgs/:o` → just `orgSlug`;
`/orgs/:o/projects/:p` adds `projectSlug`; the full triple adds
`envSlug`. Non-tenant paths get an empty scope so loaders can still
enrich.
- **`_app/route.tsx`** — already calls `requireUser`. Adds
`tenantContext.enrich({ userId: user.id })` for every authenticated
dashboard request. No new query.
- **Env layout loader (`_app.orgs.$o.projects.$p.env.$e/route.tsx`)** —
its existing `prisma.project.findFirst` gains two columns in `select`
(`externalRef`, `organization.id`). After it picks an env, calls
`tenantContext.enrich({ orgId, projectId, projectRef, envId, envType
})`. Same query, +2 columns.
- **API path (`apiBuilder.server.ts`)** — wraps every handler in
`tenantContext.run(tenantContextFromAuthEnvironment(authenticationResult.environment),
…)`. The mapper pulls `userId` from `env.orgMember?.userId` (already
selected by `authIncludeBase` — no schema change). Covers
`createLoaderApiRoute`, `createActionApiRoute`, and
`createMultiMethodApiRoute`.
- **Event processor (`sentryTenantContext.server.ts`)** — registered in
`entry.server.tsx` so it lives in the Remix bundle and shares the same
`tenantContext` ALS instance as the middleware and loaders. Stamps
whatever's present; nothing forced.
## Example events from local verification
| URL | `user.id` | Tags |
|-----|-----------|------|
| `/orgs/:o/projects/:p/env/:e/...` | real user cuid | `org_slug`,
`project_slug`, `env_slug`, `org_id`, `project_id`, `project_ref`,
`environment_id`, `env_type` |
| `/orgs/:o/settings` (non-env-scoped) | real user cuid | `org_slug`
only |
| API request with `orgMember` | `orgMember.userId` | full tenant set |
| API request without `orgMember` | (unset) | full tenant set |
## Trade-offs
1. On env-scoped pages, errors that fire before the env layout loader's
enrich callback runs get slugs + `user.id` but not the tenant IDs /
`env_type`. Realistic errors deep in async work get the full set. (Same
race as before, narrower window now that slugs/`user.id` are populated
up-front by the middleware and `_app` enrich.)
2. API requests where the environment has no `orgMember` get tenant tags
but no `user.id`. Those events still show in the issue but don't
contribute to "Users Impacted".
## Out of scope (deferred)
Background workers (`redis-worker`, `schedule-engine`) and socket
handlers. Those entry points don't set `tenantContext.run` yet — their
events ship without tenant attribution until each is wired in a
follow-up.
## Tests
31 unit tests across 4 files. New tests notably cover:
- `parseTenantPath`: org-only, org+project, and full-triple URL
variants.
- `tenantContext.enrich`: in-place patch, no-op outside `run()`,
concurrent-scope isolation, empty-scope + enrich pattern (for non-tenant
pages).
- `tenantContextFromAuthEnvironment`: with and without `orgMember` —
verifies the API path's `user.id` mapping.
- `addTenantContextToEvent`: empty scope, userId-only, slugs-only, full
enrichment, conditional tag emission, preservation of prior `event.user`
fields.
## Test plan
- [ ] `pnpm run typecheck --filter webapp`
- [ ] `pnpm run test --filter webapp -- test/tenantContext.test.ts
test/sentryTenantContext.test.ts test/tenantContextResolver.test.ts
test/tenantContextFromAuthEnvironment.test.ts`
- [ ] Local manual: with `SENTRY_DSN` set, hit a dashboard URL and an
API route, confirm the captured events carry `user.id` + the expected
tag set in Sentry.
- [ ] After ship: confirm "Users Impacted" on a real Sentry issue
reflects distinct users (not tenants).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Added `OrganizationDataStore` which allows orgs to have data stored in
specific separate services.
For now this is just used for ClickHouse. When using ClickHouse we get a
client for the factory and pass in the org id.
Particular care has to be made with two hot-insert paths:
1. RunReplicationService
2. OTLPExporter
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Claude <noreply@anthropic.com>
(`OBJECT_STORE_BASE_URL`) and a named protocol provider
(`OBJECT_STORE_DEFAULT_PROTOCOL=s3`), chat.agent session snapshot writes
landed in the named provider but reads fell through to the default — so
the recovery boot couldn't find the snapshot it had just written.
After a mid-stream cancel, the missing snapshot triggered a fallback
replay path that dropped the user's follow-up message, leaving the chat
stuck in `submitted` indefinitely.
Fix:
- New `/api/v1/sessions/:id/snapshot-url` route handles PUT + GET
symmetrically — both prefix unprefixed keys with
`OBJECT_STORE_DEFAULT_PROTOCOL` so they always round-trip through the
same store.
- `Session.chatSnapshotStoragePath` persists the resolved URI on first
write so future protocol changes don't strand existing snapshots.
Reads prefer the stored URI and fall back to the computed default for
pre-column sessions.
- SDK calls `createChatSnapshotUploadUrl` / `getChatSnapshotUrl`; the
generic v1/v2 packets endpoints are unchanged.
## Test plan
- [x] Configure local with two providers (R2 default + MinIO `s3` named)
and `OBJECT_STORE_DEFAULT_PROTOCOL=s3`.
- [x] Reproduce hang: send a message, cancel mid-stream, send another —
without the fix it hangs in `submitted`; with the fix it streams.
- [x] Snapshot lands in the `s3`-protocol bucket and
`Session.chatSnapshotStoragePath` is set after first write.
- [x] SDK unit tests pass; webapp typecheck passes.
Env-var lookups via `GET
/api/v1/projects/:projectRef/envvars/:slug/:name` run a Prisma
`findMany` on `EnvironmentVariableValue` filtered by `environmentId` +
`isSecret`. The only existing indexes are the primary key and a unique
on `(variableId, environmentId)`, so `environmentId` is never the
leading column — the planner falls back to a Parallel Seq Scan over the
whole table to find what is, in practice, a handful of rows per
environment.
Two changes:
- Add a btree index on `EnvironmentVariableValue(environmentId)` so the
planner switches to an index scan. The composite `(variableId,
environmentId)` unique stays in place; the new index is purely additive.
- Route the `findMany` inside `getEnvironmentWithRedactedSecrets`
through the read replica via a new `replicaClient` constructor param on
the repository (defaulting to `$replica`, mirroring how `prismaClient`
defaults to `prisma`). Writes and read-after-write methods stay on the
primary.
## Test plan
- [ ] `pnpm run typecheck --filter webapp`
- [ ] Confirm `EXPLAIN` plan flips from Parallel Seq Scan to an index
scan
- [ ] Existing env-var route tests still pass
## Summary
- Dashboard loaders for runs / sessions / batches / schedule-detail
threw bare `Error("X not found")` when a slug didn't resolve. Remix
surfaces this as a 500 and Sentry captures it via auto-instrumentation,
producing ongoing noise from real users following stale preview-branch
or deleted-resource links (the URLs in those Sentry events all carry
`?_data=routes/...`, i.e. client-side revalidation, not full-page
navigation).
- Added a `throwNotFound(statusText)` helper in
`app/utils/httpErrors.ts` that throws a Response with status 404,
matching the established pattern in sibling routes (agents, alerts,
bulk-actions, etc.).
- Migrated 5 loader sites to `throwNotFound` (4× "Environment not
found", 1× "Schedule not found").
- Migrated 1 loader site (`runs._index` project branch) to
`redirectWithErrorMessage("/", request, "Project not found")` to match
the pre-existing convention used by every other dashboard route's
project-not-found branch.
- Intentionally **not** touched: bare `throw new Error("X not found")`
inside `resources.*` action routes (sit inside try/catch blocks that
already redirect with a flash message), the invariant assertion in
`vercel.connect.tsx`, and the admin config check in
`admin.api.v1.runs-replication.backfill.ts`.
## Where the fix is visible
Normal browser navigation to these URLs doesn't reach the buggy loaders
— the parent env-layout
(`_app.orgs.$organizationSlug.projects.$projectParam.env.$envParam/route.tsx`)
already filters missing envs/projects and redirects/404s before the
child loader runs. The bug fires exclusively when Remix calls a single
child loader via `?_data=routes/...`, which happens during client-side
navigation or `useRevalidator`. That matches every Sentry event URL.
## Test plan
- [x] Unit test for the new helper —
`apps/webapp/test/httpErrors.test.ts`
- [x] `pnpm run typecheck --filter webapp` clean
- [x] Manual verification via Playwright on `main` vs this branch (6
cases): main returns 500 for each defective `_data` URL; branch returns
404 or 204 + `X-Remix-Redirect` as designed
- [x] Verified user-visible 404 catch boundary on `schedules/<missing>`
(the one case reachable via normal nav)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
The S2 access-token cache key was `${basin}:${streamPrefix}` — purely
server-derived but blind to the **scope/ops list** hardcoded one method
away. When the ops list changes in code (e.g. #3644 added `trim` so
`chat.agent`'s per-turn trim chain can issue `AppendRecord.trim()`),
pre-deploy tokens still in cache get returned to SDK callers for up to
the token's TTL (24h default), surfacing as `Operation not permitted`
403s on any op outside the old scope.
## Fix
Lift the ops list to a module constant and fold its sorted-join
fingerprint into the cache key:
```ts
const S2_TOKEN_OPS = ["append", "create-stream", "trim"] as const;
const S2_TOKEN_OPS_FINGERPRINT = [...S2_TOKEN_OPS].sort().join(",");
// in getS2AccessToken
const cacheKey = `${this.basin}:${this.streamPrefix}:${S2_TOKEN_OPS_FINGERPRINT}`;
// in s2IssueAccessToken
scope: { /* ... */ ops: [...S2_TOKEN_OPS], /* ... */ }
```
The fingerprint is derived from the single source of truth, so any
future scope change auto-invalidates without anyone remembering to bump
a literal version. The Unkey L1 (in-memory LRU) and L2 (Redis) layers
share the same key derivation, so both reset together on the next deploy
with no manual cache busting.
## Test plan
- [ ] `pnpm run typecheck --filter webapp`
- [ ] Run a multi-turn `chat.agent` chat via `references/ai-chat` and
confirm no `chat.agent: trim failed; will retry next turn` warn span
fires across turn-completes.
## Summary
Companion to #3536, which patched routes that already had a leaking
`catch (e) { return json({error: e.message}, 500) }`. That pattern can't
reach routes which have no catch in the first place — when those throw,
Remix's default error path serializes `error.message` into the response
body, and the SDK then wraps the leaked string as `TriggerApiError`.
Across 28 raw api.v1 loaders/actions plus one dashboard polling
endpoint, each handler now:
- Wraps its body in `try { ... } catch (error) { ... }`.
- Re-throws `Response` instances so auth helpers' `throw json(...)` /
`throw redirect(...)` pass through unchanged.
- Logs non-Response errors via `logger.error` so server-side visibility
is preserved.
- Returns a generic body — `{"error": "Internal Server Error"}` 500 for
raw API routes, or `{ changelogs: [] }` 200 for the polling widget
(degrade silently across transient blips; the consumer hook already
coped with empty payloads).
For six routes where #3536 left an inner try/catch covering only a
service call (`alertChannels`, `batches.results`,
`deployments.finalize`, `deployments.background-workers`,
`deployments.promote`, `projects.background-workers`): an outer
try/catch is added so auth/parsing failures are also sanitized. Inner
typed-error handling (`ServiceValidationError` → 422 with message, etc.)
is preserved exactly.
For two routes whose existing catch returned 400 + `error.message`
(`api.v1.authorization-code`, `api.v1.orgs.\$orgParam.projects` action):
the body is sanitized to a generic per-route string. **Status code stays
400** — clients that key on the 4xx/5xx distinction (and the SDK's
no-retry-on-4xx behavior) are unaffected.
## Test plan
- [x] \`pnpm run typecheck --filter webapp\`
- [x] Per-route synthetic-throw probe: inject \`throw new
Error("SYNTHETIC ...")\` at the top of each catch'd try, curl the route
with a dummy bearer, confirm the response body is the generic shape and
that the synthetic message lands server-side via \`logger.error\`. 29
routes verified.
- [x] Real-P1001 probe on the envvars loader: \`docker stop database\`
mid-flight, confirm response is generic 500 (not the leaked Prisma
message).
- [x] Sampled legitimate 4xx/2xx paths across each pattern variant
(naked-wrap, partial-expanded, 400-preserved) to confirm the wraps don't
interfere with normal control flow.
## Summary
- Introduce the Mollifier: a Redis-backed buffer for `trigger()` API
calls during traffic spikes, with a per-env trip evaluator and a drainer
ack-loop.
- Phase 1 is dual-write monitoring — every mollified trigger is buffered
to Redis AND continues to `engine.trigger`. No customer-facing behaviour
change.
- Telemetry events: `mollifier.would_mollify`, `mollifier.buffered`,
`mollifier.drained`, plus the `mollifier.decisions` counter.
- Gated behind a feature flag (default off).
## Test plan
- [x] `pnpm run test --filter @trigger.dev/redis-worker`
- [x] `pnpm run test --filter webapp -- mollifier`
- [x] Manual: with flag off, no behaviour change vs main
- [x] Manual: with flag on + threshold lowered, observe
`mollifier.buffered` + `mollifier.drained` log pairs with matching
`runId`
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
## Summary
Long-running chat agents were filling `session.out` forever — every
`chat.agent` turn appended to the same S2 stream with no trim, and the
Sessions dashboard re-streamed the entire history from `seq_num=0` on
every page load. After this change the agent appends an S2 `trim`
command record after each `trigger:turn-complete`, pointing back at the
previous turn-complete's seq_num. `session.out` stays roughly one turn
long at steady state, regardless of session age.
`trigger:turn-complete` and `trigger:upgrade-required` move from
`chunk.type`-shaped data records into header-form control records under
a uniform `trigger-control` namespace. Built-in transports
(`TriggerChatTransport`, `AgentChat`, the dashboard's `AgentView`)
handle the new shape transparently. Custom transports need a one-line
filter on the `trigger-control` header — see the rewritten "Records on
session.out" section in the client-protocol docs.
The Sessions detail page in the dashboard fetches the agent's per-turn
S3 snapshot via a presigned URL and seeds the transcript view, then
SSE-tails from the snapshot's `lastOutEventId`. Bandwidth and
time-to-first-render scale with unread turns instead of session
lifetime.
Resume contract is now explicit: single-turn-boundary resume always
works (the prior turn-complete is still on the stream), the S2 trim is
eventually consistent over 10-60s, and multi-turn-away resume falls back
to a snapshot reload.
## Summary
The PUT handler at `/realtime/v1/streams/:runId/:target/:streamId` ran
`taskRun.update({ realtimeStreams: { push: streamId } })` on every call,
even when the `streamId` was already present. SDK call patterns that
re-initialize the same stream key on every chunk produce a per-write row
UPDATE, duplicate entries pile up in the array, and the row-lock + TOAST
rewrite cost grows unbounded on long-running stream sessions.
## Fix
Mirror the sibling append handler: read the array first and only push
when the `streamId` isn't already present. Identical behavior for
first-time stream creation; repeat creates short-circuit to a single
indexed read. The dashboard's per-run stream listing keeps working
because the first create still records the entry.
## Test plan
- [ ] A fresh PUT for a new `(run, streamId)` adds the entry to the
array
- [ ] A repeat PUT for the same pair leaves the array unchanged
- [ ] 404 is returned when the run doesn't exist; 400 when the run is
completed
## Summary
Adds a Region column and Region filter (under More filters) to the runs
list dashboard, the same filter on the public runs list API
(`filter[region]`), and a matching `region` input on the MCP `list_runs`
tool. Each run's executing region is also surfaced as a new optional
`region` field on the runs list and run retrieve responses, populated
from the worker instance group's `masterQueue` identifier.
Useful when you run tasks across multiple regions and want to slice the
runs list — or your existing run-querying scripts — by where the run
actually executed.
## Design
The filter value in the URL / API is the `masterQueue` identifier (the
same string already persisted on `TaskRun` and replicated to ClickHouse
as `worker_queue`), so the query just becomes `worker_queue IN (...)`
with no server-side translation. The Region dropdown options come from a
new resource loader backed by `RegionsPresenter`, which now also exposes
`masterQueue` alongside the existing region metadata.
```ts
// public API
const runs = await runs.list({ region: ["us-east-1", "eu-west-1"] });
// each item: { id, status, ..., region?: "us-east-1" }
```
```ts
// MCP
list_runs({ environment: "prod", region: "us-east-1" })
```
## Summary
The trigger-task hotpath used to early-return without a DB query when a
caller passed both a queue override and a per-trigger TTL — the hottest
configuration on the trigger API. Adding `triggerSource` to the resolver
so the runs-list "Source" filter could distinguish STANDARD / SCHEDULED
/
AGENT runs removed those early-returns, costing +2 DB queries per
trigger
on non-locked calls and +1 on locked calls.
This change caches `BackgroundWorkerTask` metadata (`ttl`,
`triggerSource`,
`queueId`, `queueName`) in Redis so the resolver can satisfy every
caller
configuration with a single `HGET` on the warm path. PG fallback on miss
back-fills the cache.
Follow-up to #3542.
## Design
Two key spaces:
- `task-meta:env:{envId}` — the "current worker" view, refreshed at
every
deploy promotion. 24h safety TTL.
- `task-meta:by-worker:{workerId}` — used for `lockToVersion` triggers.
Immutable post-create. 30d sliding TTL so historical workers age out.
Cache writes use Lua scripts via `defineCommand` so `DEL` + `HSET` +
`EXPIRE` land atomically — concurrent readers never see the empty
intermediate state of a naive pipeline. Read-path back-fill uses
single-field upserts so concurrent back-fills don't wipe each other's
siblings.
The cache lives behind its own `TASK_META_CACHE_REDIS_*` env-var prefix
that falls back to the default `REDIS_*` set, so operators can route the
cache to a dedicated Redis instance if they want.
The service/instance file split (`taskMetadataCache.server.ts` for the
pure class, `taskMetadataCacheInstance.server.ts` for the env-wired
singleton) mirrors the existing `runsReplicationService` /
`runsReplicationInstance` pattern.
## Test plan
- [ ] `pnpm run typecheck --filter webapp`
- [ ] `pnpm run test ./test/engine/triggerTask.test.ts --run` — 8
existing tests untouched + 5 new tests covering warm cache, cold
miss with back-fill, queue + ttl path, by-worker vs env keyspace,
and the promotion cache write
- [ ] End-to-end against a dev worker: registering writes both keyspaces
with the expected TTLs, and `redis-cli HGETALL
"tr:task-meta:env:<envId>"`
returns the cached entries
## Benchmark
Measured `DefaultQueueManager.resolveQueueProperties` against a real
Postgres + Redis (vitest `containerTest`, single-host docker). 500
sequential calls and 2,000 parallel calls (concurrency=50) per scenario,
request shaped as `{ taskId, queue: "bench-queue", ttl: "5m" }` — the
hot path this PR restores.
```
sequential (one in flight at a time):
[noop cache (baseline)] n=500 mean=1.423ms p50=1.394ms p95=1.735ms p99=2.629ms max=11.100ms
[redis cache, cold ] n=500 mean=1.346ms p50=1.283ms p95=1.688ms p99=2.463ms max=5.058ms
[redis cache, warm ] n=500 mean=0.084ms p50=0.078ms p95=0.105ms p99=0.156ms max=1.129ms
speedup (warm vs baseline, sequential): 16.95x
parallel (concurrency=50):
[noop cache (baseline)] n=2000 mean=10.069ms p50=8.850ms p95=14.718ms p99=31.887ms total=405ms ops/s=4,940
[redis cache, warm ] n=2000 mean=0.614ms p50=0.568ms p95=1.189ms p99=1.432ms total=25ms ops/s=80,389
throughput speedup (warm vs baseline, parallel): 16.27x
```
Read:
- **Warm cache cuts resolver latency 17×** at p50 — from ~1.4 ms to ~78
µs per call.
- **Cold cache is on par with baseline** — the extra `HGET` miss adds
<50 µs against the two Postgres queries that follow, so the worst case
is not worse than today.
- **Under burst load (50 concurrent triggers)**, the baseline's p99
jumps to ~32 ms as Postgres connections queue up; warm stays at ~1.4 ms.
The cache moves the saturation point from ~5k ops/s (PG pool) to ~80k
ops/s (single-client Redis pipelining).
Caveats: single-host docker, local Postgres + Redis, resolver-only
measurement (excludes the rest of the trigger transaction). Prod adds
region-local Redis RTT (~0.3–0.8 ms) which shifts warm absolute numbers
up but keeps the ratio intact.
## Summary
A "Google auth conflict" Sentry alert fires whenever a user signs in via
Google whose Google account is linked to one user row but whose
Google-provided email is now on a *different* user row. The handler in
`apps/webapp/app/models/user.server.ts:236` already does the right thing
— it returns the existing auth-linked user and skips the update path so
neither row gets mutated — but it logs the situation with
`logger.error`, which routes to Sentry as an exception and pages the
on-call channel.
There's no exception to chase here: the branch is the intended outcome
for a known data shape (user changed their email on one account after
originally signing up via Google on another). Downgrading the call to
`logger.warn` keeps the diagnostic record in our logs (with all the same
context fields — email, both user IDs, authIdentifier) but stops it
firing the production error alert.
## Change
- `logger.error` → `logger.warn` for the conflict branch in
`findOrCreateGoogleUser`. Context payload is unchanged.
## Test plan
- [x] Typecheck only — there's no behavioural change to test, the log
level is the entire diff.
## Summary
A chat-aware run inspector and a `/playground` UI for testing
`chat.agent` tasks interactively. Builds on #3543's runtime.
## Design
The run inspector grows a new tab that renders the conversation chain
for any `chat.agent`-kind run. It subscribes to the run's session
streams, threads chat parts through a per-message renderer, and uses a
shared markdown + Shiki component for code highlighting (also used by
the test-payload panel).
The playground is a standalone `/playground` route that lets you drive a
deployed chat agent from the dashboard — pick a task, send messages,
watch tool calls render, and see span detail on every turn. The matching
`/agents` list view shows all deployed agents in the project.
Concurrent `POST /api/v1/deployments` requests for the same environment
race on the `WorkerDeployment(environmentId, version)` unique
constraint. Both requests read the same latest deployment via
`findFirst`, compute the same next version via
`calculateNextBuildVersion`, and both attempt
`prisma.workerDeployment.create()` — one wins, the other crashes with
Prisma `P2002`. The bug is a classic TOCTOU between the version read and
the version write; it's been latent since the version-assignment logic
was first added but only fires when two deploys land within milliseconds
of each other (CI matrices, retried CLI calls, webhook-triggered
redeploys).
## Approach
Extracts the version assignment + create into a small helper
`createDeploymentWithNextVersion`
(`apps/webapp/app/v3/services/initializeDeployment/createDeploymentWithNextVersion.server.ts`).
The helper retries on `P2002 (environmentId, version)` up to 5 times
with randomised 5–50ms jitter so N concurrent racers don't loop in
lockstep. Each attempt re-reads the latest version, recomputes via
`calculateNextBuildVersion`, and re-runs the caller's `buildData`
callback so version-dependent fields (image ref tag, friendlyId) are
always consistent with the version actually persisted. A `logger.warn`
fires per collision so the retry rate is observable in production logs.
When retries are exhausted, the helper throws a dedicated
`DeploymentVersionCollisionError` carrying `environmentId`, `attempts`,
and `lastAttemptedVersion`, with the original
`PrismaClientKnownRequestError` attached as `cause`. Sentry walks the
`cause` chain natively, so contention exhaustion shows up as a
distinguishable wrapper exception linked to the underlying `P2002`
rather than a generic unique-constraint violation that looks identical
to every other duplicate-key bug.
The behavioural change is limited to "catch P2002 and retry instead of
crashing." The image ref computation stays inside the builder callback
(same call site as before the refactor), so ECR / non-ECR behaviour, S2
stream creation order, and all downstream side effects are unchanged.
## Non-goals
- No new database migrations, no schema changes, no isolation-level /
locking changes. A serialisable transaction or advisory lock would also
fix this; retry-on-conflict is the smaller change that keeps the
existing version-allocation logic intact.
- Does not touch the analogous `calculateNextBuildVersion` call in
`createBackgroundWorker.server.ts`, which likely has the same race shape
against `BackgroundWorker`'s unique constraint — flagged as a follow-up.
## Test plan
- [x] `pnpm run typecheck --filter webapp` passes (no new errors in the
modified files).
- [x] Three real-Postgres tests in
`apps/webapp/test/createDeploymentWithNextVersion.test.ts` via
`containerTest`:
- 5 concurrent calls all produce distinct, persistable versions
(`Set(versions).size === concurrency`). The naive read-then-create
version of the helper fails this test with the exact same `P2002` seen
in production; the retry version passes.
- Non-`P2002` errors raised from the `buildData` callback propagate
immediately without retry, builder invoked exactly once.
- With `maxRetries: 0`, concurrent racers surface the wrapped
`DeploymentVersionCollisionError` (not a raw `P2002`); `environmentId`,
`attempts`, `lastAttemptedVersion` are populated and `error.cause.code
=== "P2002"`.
- [x] Existing `apps/webapp/test/getDeploymentImageRef.test.ts` still
green (the file was untouched in the final diff).
## Follow-ups (not in this PR)
- `createBackgroundWorker.server.ts` likely has the same TOCTOU shape
against its background-worker version unique constraint — should use the
same helper.
- Sentry visibility check: confirm `error.cause` chain renders as a
linked exception in the Sentry UI when the wrapped error fires (requires
a sandboxed triggering of the exhaustion path).
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds the chat.agent({...}) task definition (server runtime) and the
browser-side TriggerChatTransport + AgentChat that drives it from a
React or Next.js app. The runtime sits on top of the Sessions primitive
and handles the durable conversational task lifecycle.
Server runtime:
- chat.agent({...}) — session-aware task definition
- Lifecycle hooks: onChatStart, onTurnStart, onTurnComplete, onAction,
onValidateMessages, hydrateMessages
- chat.history read primitives for HITL flows
- chat.local, chat.headStart, chat.handover, oomMachine
- Delta-only wire + S3 snapshot reconstruction at run boot
- Actions are no longer turns
Browser transport:
- TriggerChatTransport (ai-sdk Transport): delta-only wire sends,
SSE reconnection with lastEventId resume, stop/abort cleanup,
dynamic accessToken refresh
- AgentChat: direct programmatic API
- useTriggerChatTransport (React hook)
- chat-tab-coordinator: cross-tab leader election
Includes the chat-agent, chat-agent-delta-wire-snapshots,
chat-history-read-primitives, chat-head-start, chat-actions-no-turn,
chat-session-attributes, agent-skills, and mock-chat-agent-test-harness
changesets.
Dan