Release 3.0.0-beta.45

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

.changeset/afraid-sheep-joke.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fixes an issue with scoped packages in additionalPackages option

.changeset/angry-eagles-trade.md

@@ -0,0 +1,9 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+- Fix additionalFiles that aren't decendants
+- Stop swallowing uncaught exceptions in prod
+- Improve warnings and errors, fail early on critical warnings
+- New arg to --save-logs even for successful builds

.changeset/beige-pears-explode.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Add an e2e suite to test compiling with v3 CLI.

.changeset/beige-pens-dance.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+v3 CLI update command and package manager detection fix

.changeset/big-tomatoes-deliver.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Hoist uncaughtException handler to the top of workers to better report error messages

.changeset/blue-pumas-whisper.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/shopify": patch
+---
+
+improved error messages when a shopify webhook fails to register

.changeset/brown-spies-burn.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+cli v3: increase otel force flush timeout to 30s from 500ms

.changeset/clever-apes-collect.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Fix error stack traces

.changeset/clever-carrots-travel.md

@@ -0,0 +1,7 @@
+---
+"@trigger.dev/sdk": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Implement task.onSuccess/onFailure and config.onSuccess/onFailure

.changeset/clever-donkeys-hunt.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Remove "log" Log Level, unify log and info messages under the "info" log level

.changeset/cool-comics-burn.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fixing missing logs when importing client @opentelemetry/api

.changeset/dull-mangos-press.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Strip out server-only package from worker builds

.changeset/eleven-paws-join.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Remove the env var check during deploy (too many false negatives)

.changeset/famous-boats-tease.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Better handle issues with resolving dependency versions during deploy

.changeset/fast-colts-relax.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Increase dev worker timeout

.changeset/five-toes-destroy.md

@@ -0,0 +1,7 @@
+---
+"@trigger.dev/sdk": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+v3: Environment variable management API and SDK, along with resolveEnvVars CLI hook

.changeset/friendly-walls-repair.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Add sox and audiowaveform binaries to worker images

.changeset/funny-swans-destroy.md

@@ -0,0 +1,9 @@
+---
+"@trigger.dev/sdk": patch
+---
+
+Make msw a normal dependency (for now) to fix Module Not Found error in Next.js.
+
+It turns out that webpack will "hoist" dynamically imported modules and attempt to resolve them at build time, even though it's an optional peer dep:
+
+https://x.com/maverickdotdev/status/1782465214308319404

.changeset/gorgeous-gorillas-compete.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+Add runs.retrieve management API method to get info about a run by run ID

.changeset/hot-buckets-behave.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Support triggering tasks with non-URL friendly characters in the ID

.changeset/hot-fishes-retire.md

@@ -0,0 +1,5 @@
+---
+trigger.dev: patch
+---
+
+Fix TypeScript inclusion in tsconfig.json for `cli-v3 init`

.changeset/hot-wasps-sin.md

@@ -0,0 +1,7 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+"@trigger.dev/sdk": patch
+---
+
+v3: Usage tracking

.changeset/itchy-chairs-itch.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/sdk": patch
+---
+
+Fix for calling trigger and passing a custom queue

.changeset/lazy-files-lay.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Increased the timeout when canceling a checkpoint to 31s (to match the timeout on the server)

.changeset/little-crabs-cross.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Fix issues with consecutive waits

.changeset/long-fireants-search.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Adding some additional telemetry during deploy to help debug issues

.changeset/lovely-drinks-flash.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Add the "log" level back in as an alias to "info"

.changeset/mighty-flowers-train.md

@@ -0,0 +1,83 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Adds support for `emitDecoratorMetadata: true` and `experimentalDecorators: true` in your tsconfig using the [`@anatine/esbuild-decorators`](https://github.com/anatine/esbuildnx/tree/main/packages/esbuild-decorators) package. This allows you to use libraries like TypeORM: 
+
+```ts orm/index.ts
+import "reflect-metadata";
+import { DataSource } from "typeorm";
+import { Entity, Column, PrimaryColumn } from "typeorm";
+
+@Entity()
+export class Photo {
+  @PrimaryColumn()
+  id!: number;
+
+  @Column()
+  name!: string;
+
+  @Column()
+  description!: string;
+
+  @Column()
+  filename!: string;
+
+  @Column()
+  views!: number;
+
+  @Column()
+  isPublished!: boolean;
+}
+
+export const AppDataSource = new DataSource({
+  type: "postgres",
+  host: "localhost",
+  port: 5432,
+  username: "postgres",
+  password: "postgres",
+  database: "v3-catalog",
+  entities: [Photo],
+  synchronize: true,
+  logging: false,
+});
+```
+
+And then in your trigger.config.ts file you can initialize the datasource using the new `init` option:
+
+```ts trigger.config.ts
+import type { TriggerConfig } from "@trigger.dev/sdk/v3";
+import { AppDataSource } from "@/trigger/orm";
+
+export const config: TriggerConfig = {
+  // ... other options here
+  init: async (payload, { ctx }) => {
+    await AppDataSource.initialize();
+  },
+};
+```
+
+Now you are ready to use this in your tasks:
+
+```ts
+import { task } from "@trigger.dev/sdk/v3";
+import { AppDataSource, Photo } from "./orm";
+
+export const taskThatUsesDecorators = task({
+  id: "taskThatUsesDecorators",
+  run: async (payload: { message: string }) => {
+    console.log("Creating a photo...");
+
+    const photo = new Photo();
+    photo.id = 2;
+    photo.name = "Me and Bears";
+    photo.description = "I am near polar bears";
+    photo.filename = "photo-with-bears.jpg";
+    photo.views = 1;
+    photo.isPublished = true;
+
+    await AppDataSource.manager.save(photo);
+  },
+});
+```

.changeset/mighty-parrots-sin.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+v3: Remove aggressive otel flush timeouts in dev/prod

.changeset/modern-stingrays-end.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+v3: Trigger delayed runs and reschedule them

.changeset/nasty-jars-pump.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fixes an issue that was treating v2 trigger directories as v3

.changeset/nervous-planets-sparkle.md

@@ -0,0 +1,9 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+- Improve non-zero exit code error messages
+- Detect OOM conditions within worker child processes
+- Internal errors can have optional stack traces
+- Docker provider can be set to enforce machine presets

.changeset/new-pants-beg.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Fix issue when using SDK in non-node environments by scoping the stream import with node:

.changeset/nice-bulldogs-turn.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Better handle uncaught exceptions

.changeset/pink-pumas-rhyme.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+Add configure function to be able to configure the SDK manually

.changeset/plenty-ducks-beam.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Output stderr logs on dev worker failure

.changeset/polite-pears-grow.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/sdk": patch
+---
+
+Use global setTimeout to ensure cross-runtime support

.changeset/polite-pots-walk.md

@@ -0,0 +1,7 @@
+---
+"trigger.dev": patch
+---
+
+Add e2e fixtures corresponding to past issues
+Implement e2e suite parallelism
+Enhance log level for specific e2e suite messages

.changeset/polite-rockets-matter.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fix dev CLI output when not printing update messages

.changeset/pre.json

@@ -44,53 +44,133 @@
     "@trigger.dev/yalt": "2.3.18"
   },
   "changesets": [
+    "afraid-sheep-joke",
+    "angry-eagles-trade",
+    "beige-pears-explode",
+    "beige-pens-dance",
+    "big-tomatoes-deliver",
+    "blue-pumas-whisper",
     "breezy-gorillas-mate",
+    "brown-spies-burn",
     "chilled-hornets-move",
     "clean-pianos-listen",
+    "clever-apes-collect",
+    "clever-carrots-travel",
+    "clever-donkeys-hunt",
+    "cool-comics-burn",
     "cool-glasses-bake",
     "cuddly-feet-approve",
     "dry-walls-check",
+    "dull-mangos-press",
     "eight-pumas-float",
+    "eleven-paws-join",
+    "famous-boats-tease",
+    "fast-colts-relax",
     "few-students-share",
+    "five-toes-destroy",
+    "friendly-walls-repair",
+    "funny-swans-destroy",
+    "gorgeous-gorillas-compete",
     "green-bags-wink",
+    "hot-buckets-behave",
+    "hot-fishes-retire",
+    "hot-wasps-sin",
+    "itchy-chairs-itch",
     "khaki-apricots-design",
     "khaki-poems-lay",
     "late-icons-lie",
     "late-steaks-behave",
+    "lazy-files-lay",
     "lemon-jobs-repair",
     "light-bulldogs-press",
     "light-dragons-complain",
+    "little-crabs-cross",
+    "long-fireants-search",
     "loud-actors-remember",
+    "lovely-drinks-flash",
     "many-ligers-pump",
     "mighty-camels-joke",
+    "mighty-flowers-train",
+    "mighty-parrots-sin",
+    "modern-stingrays-end",
+    "nasty-jars-pump",
+    "nervous-planets-sparkle",
+    "new-pants-beg",
     "new-rivers-tell",
+    "nice-bulldogs-turn",
     "ninety-pets-travel",
     "odd-poets-own",
+    "pink-pumas-rhyme",
+    "plenty-ducks-beam",
     "polite-ducks-switch",
+    "polite-pears-grow",
+    "polite-pots-walk",
+    "polite-rockets-matter",
     "poor-flowers-cross",
+    "purple-garlics-shop",
+    "purple-spiders-care",
+    "rare-lamps-promise",
     "rare-roses-float",
     "real-planets-stare",
+    "rich-kangaroos-unite",
+    "rotten-beers-refuse",
     "rotten-dryers-exercise",
+    "rude-toys-compare",
+    "selfish-ducks-sort",
+    "serious-hats-rest",
     "shaggy-spoons-taste",
     "sharp-emus-compare",
     "sharp-zebras-serve",
     "shiny-coats-cry",
+    "silly-buses-obey",
+    "silly-forks-kiss",
     "silly-suits-switch",
+    "silver-doors-juggle",
+    "six-ligers-exist",
+    "six-rats-hunt",
+    "sixty-insects-watch",
+    "slow-buses-own",
+    "slow-kiwis-hide",
+    "slow-sloths-retire",
     "smart-needles-move",
     "smart-olives-eat",
+    "sour-pugs-teach",
+    "spicy-frogs-remain",
     "spicy-lamps-smoke",
+    "spicy-terms-bow",
     "strange-ghosts-matter",
+    "strange-sheep-pull",
     "strong-lemons-add",
+    "strong-owls-know",
+    "strong-phones-smoke",
+    "stupid-adults-sniff",
     "stupid-bulldogs-applaud",
+    "sweet-ducks-remember",
     "sweet-lizards-press",
     "swift-dragons-peel",
     "tall-bees-wave",
+    "tall-masks-repeat",
+    "tame-apricots-clap",
     "tame-guests-know",
+    "tender-moose-tell",
     "tender-oranges-rhyme",
+    "tender-turkeys-compete",
+    "thick-carrots-sneeze",
+    "thin-parents-heal",
+    "thirty-islands-kiss",
     "tidy-balloons-suffer",
     "tidy-dryers-sleep",
+    "tidy-tomatoes-explain",
     "tiny-doors-type",
     "tiny-elephants-scream",
-    "tricky-bulldogs-heal"
+    "tricky-bulldogs-heal",
+    "tricky-keys-attack",
+    "tricky-ladybugs-unite",
+    "twelve-knives-notice",
+    "two-pumas-wait",
+    "violet-clocks-notice",
+    "warm-olives-provide",
+    "warm-planes-taste",
+    "young-snails-sell"
   ]
 }

.changeset/purple-garlics-shop.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Add typescript as a dependency so the esbuild-decorator will work even when running in npx

.changeset/purple-spiders-care.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Await file watcher cleanup in dev

.changeset/rare-lamps-promise.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Add a postInstall option to allow running scripts after dependencies have been installed in deployed images

.changeset/rich-kangaroos-unite.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fix a bug where revoking the CLI token would prevent you from ever logging in again with the CLI.

.changeset/rotten-beers-refuse.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Add git to prod worker image which fixes private package installs

.changeset/rude-toys-compare.md

@@ -0,0 +1,7 @@
+---
+"@trigger.dev/core-apps": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Capture and display stderr on index failures

.changeset/selfish-ducks-sort.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"trigger.dev": patch
+---
+
+Lock SDK and CLI deps on exact core version

.changeset/serious-hats-rest.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+v3: [prod] force flush timeout should be 1s

.changeset/silly-buses-obey.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Add callback to checkpoint created message

.changeset/silly-forks-kiss.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+v3: Copy over more of the project's package.json keys into the deployed package.json (support for custom config like zenstack)

.changeset/silver-doors-juggle.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Make deduplicationKey required when creating/updating a schedule

.changeset/six-ligers-exist.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Improve the display of non-object return types in the run trace viewer

.changeset/six-rats-hunt.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Improved ESM module require error detection logic

.changeset/sixty-insects-watch.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Set the deploy timeout to 3mins from 1min

.changeset/slow-buses-own.md

@@ -0,0 +1,8 @@
+---
+"@trigger.dev/sdk": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+"@trigger.dev/cli": patch
+---
+
+Move to our global system from AsyncLocalStorage for the current task context storage

.changeset/slow-kiwis-hide.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+v3: vercel edge runtime support

.changeset/slow-sloths-retire.md

@@ -0,0 +1,7 @@
+---
+"@trigger.dev/core-apps": patch
+"@trigger.dev/core": patch
+---
+
+- Fix uncaught provider exception
+- Remove unused provider messages

.changeset/sour-pugs-teach.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+v3: fix otel flushing causing CLEANUP ack timeout errors by always setting a forceFlushTimeoutMillis value

.changeset/spicy-frogs-remain.md

@@ -0,0 +1,10 @@
+---
+"trigger.dev": patch
+---
+
+- Prevent downgrades during update check and advise to upgrade CLI
+- Detect bun and use npm instead
+- During init, fail early and advise if not a TypeScript project
+- During init, allow specifying custom package manager args
+- Add links to dev worker started message
+- Fix links in unsupported terminals

.changeset/spicy-terms-bow.md

@@ -0,0 +1,9 @@
+---
+"@trigger.dev/core-apps": patch
+"trigger.dev": patch
+---
+
+- Fix init command SDK pinning
+- Show --api-url / -a flag where needed
+- CLI now also respects `TRIGGER_TELEMETRY_DISABLED`
+- Dedicated docker checkpoint test function

.changeset/strange-sheep-pull.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+v2: Better handle recovering from platform communication errors by auto-yielding back to the platform in case of temporary API failures

.changeset/strong-owls-know.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+When a v2 run hits the rate limit, reschedule with the reset date

.changeset/strong-phones-smoke.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+fix: allow command login to read api url from cli args

.changeset/stupid-adults-sniff.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+Management SDK overhaul and adding the runs.list API

.changeset/sweet-ducks-remember.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/sdk": patch
+---
+
+v3: Adding SDK functions for triggering tasks in a typesafe way, without importing task file

.changeset/tall-masks-repeat.md

@@ -0,0 +1,9 @@
+---
+"@trigger.dev/core-apps": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+- Fix artifact detection logs
+- Fix OOM detection and error messages
+- Add test link to cli deployment completion

.changeset/tame-apricots-clap.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+v3: postInstall config option now replaces the postinstall script found in package.json

.changeset/tender-moose-tell.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Use locked package versions when resolving dependencies in deployed workers

.changeset/tender-turkeys-compete.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core": patch
+---
+
+Added timezone support to schedules

.changeset/thick-carrots-sneeze.md

@@ -0,0 +1,6 @@
+---
+"@trigger.dev/core": patch
+"@trigger.dev/sdk": patch
+---
+
+v3: Include presigned urls for downloading large payloads and outputs when using runs.retrieve

.changeset/thin-parents-heal.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Ensure @trigger.dev/sdk and @trigger.dev/core are always in the list of deployed dependencies

.changeset/thirty-islands-kiss.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fix for typo in v3 CLI login command

.changeset/tidy-tomatoes-explain.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Fix 3rd party otel propagation from breaking our Task Events data from being properly correlated to the correct trace

.changeset/tricky-keys-attack.md

@@ -0,0 +1,14 @@
+---
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+- Clear paused states before retry
+- Detect and handle unrecoverable worker errors
+- Remove checkpoints after successful push
+- Permanently switch to DO hosted busybox image
+- Fix IPC timeout issue, or at least handle it more gracefully
+- Handle checkpoint failures
+- Basic chaos monkey for checkpoint testing
+- Stack traces are back in the dashboard
+- Display final errors on root span

.changeset/tricky-ladybugs-unite.md

@@ -0,0 +1,8 @@
+---
+"@trigger.dev/sdk": patch
+"@trigger.dev/core-apps": patch
+"trigger.dev": patch
+"@trigger.dev/core": patch
+---
+
+Extracting out all the non-SDK related features from the main @trigger.dev/core/v3 export

.changeset/twelve-knives-notice.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+v3: fix missing init output in task run function when no middleware is defined

.changeset/two-pumas-wait.md

@@ -0,0 +1,6 @@
+---
+"trigger.dev": patch
+"@trigger.dev/cli": patch
+---
+
+Add support for tasks located in subdirectories inside trigger dirs

.changeset/violet-clocks-notice.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Fix jsonc-parser import

.changeset/warm-olives-provide.md

@@ -0,0 +1,5 @@
+---
+"@trigger.dev/core": patch
+---
+
+Improve handling of IPC timeouts and fix checkpoint cancellation after failures

.changeset/warm-planes-taste.md

@@ -0,0 +1,7 @@
+---
+"trigger.dev": patch
+---
+
+v3: Prevent legacy-peer-deps=true from breaking deploys
+
+When a global `.npmrc` file includes `legacy-peer-deps=true`, deploys would fail on the `npm ci` step because the package-lock.json wouldn't match the `package.json` file. This is because inside the image build, the `.npmrc` file would not be picked up and so `legacy-peer-deps` would end up being false (which is the default). This change forces the `package-lock.json` file to be created using `legacy-peer-deps=false`

.changeset/young-snails-sell.md

@@ -0,0 +1,5 @@
+---
+"trigger.dev": patch
+---
+
+Increase cleanup IPC timeout

.env.example

@@ -11,15 +11,22 @@ REMIX_APP_PORT=3030
 APP_ENV=development
 APP_ORIGIN=http://localhost:3030
 NODE_ENV=development
+CLOUD_ENV=development
+V3_ENABLED=true
 
-# Redis is used for concurrency control
-# REDIS_HOST="localhost"
-# REDIS_PORT="6379"
-# REDIS_TLS_DISABLED="true"
+# Redis is used for the v3 queuing and v2 concurrency control
+REDIS_HOST="localhost"
+REDIS_PORT="6379"
+REDIS_TLS_DISABLED="true"
+
+DEV_OTEL_EXPORTER_OTLP_ENDPOINT="http://localhost:3030/otel"
+DEV_OTEL_BATCH_PROCESSING_ENABLED="0"
 
 # OPTIONAL VARIABLES
 # This is used for validating emails that are allowed to log in. Every email that do not match this regex will be rejected.
 # WHITELISTED_EMAILS="authorized@yahoo\.com|authorized@gmail\.com"
+# Accounts with these emails will get global admin rights. This grants access to the admin UI.
+# ADMIN_EMAILS="admin@example\.com|another-admin@example\.com"
 # This is used for logging in via GitHub. You can leave these commented out if you don't want to use GitHub for authentication.
 # AUTH_GITHUB_CLIENT_ID=
 # AUTH_GITHUB_CLIENT_SECRET=

.github/workflows/e2e.yml

@@ -1,9 +1,53 @@
-name: "🧪 E2E Tests"
+name: "E2E"
 on:
   workflow_call:
+    inputs:
+      package:
+        description: The identifier of the job to run
+        default: webapp
+        required: false
+        type: string
 jobs:
-  e2e:
-    name: "🧪 E2E Tests"
+  cli-v3:
+    name: "🧪 CLI v3 tests"
+    if: inputs.package == 'cli-v3' || inputs.package == ''
+    runs-on: buildjet-8vcpu-ubuntu-2204
+    strategy:
+      fail-fast: false
+      matrix:
+        package-manager: ["npm", "pnpm", "yarn"]
+    steps:
+      - name: ⬇️ Checkout repo
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: ⎔ Setup pnpm
+        uses: pnpm/action-setup@v2.2.4
+        with:
+          version: 8.15.5
+
+      - name: ⎔ Setup node
+        uses: buildjet/setup-node@v3
+        with:
+          node-version: 20.11.1
+          cache: "pnpm"
+
+      - name: 📥 Download deps
+        run: pnpm install --frozen-lockfile --filter trigger.dev...
+
+      - name: 🔧 Build v3 cli monorepo dependencies
+        run: pnpm run build --filter trigger.dev^...
+
+      - name: 🔧 Build worker template files
+        run: pnpm --filter trigger.dev run build:workers
+
+      - name: Run E2E Tests
+        run: |
+          PM=${{ matrix.package-manager }} pnpm --filter trigger.dev run test:e2e
+  webapp:
+    name: "🧪 Webapp tests"
+    if: inputs.package == 'webapp' || inputs.package == ''
     runs-on: buildjet-16vcpu-ubuntu-2204
     steps:
       - name: 🐳 Login to Docker Hub

.github/workflows/pr_checks.yml

@@ -27,6 +27,8 @@ jobs:
     uses: ./.github/workflows/unit-tests.yml
     secrets: inherit
 
-  e2e:
-    uses: ./.github/workflows/e2e.yml
-    secrets: inherit
+  # e2e:
+  #   uses: ./.github/workflows/e2e.yml
+  #   with:
+  #     package: webapp
+  #   secrets: inherit

.github/workflows/publish-docker.yml

@@ -4,13 +4,15 @@ on:
 jobs:
   publish:
     runs-on: ubuntu-latest
+    env:
+      PRISMA_ENGINES_CHECKSUM_IGNORE_MISSING: 1
     outputs:
       version: ${{ steps.get_version.outputs.version }}
       short_sha: ${{ steps.get_commit.outputs.sha_short }}
     steps:
       - name: Setup Depot CLI
         uses: depot/setup-action@v1
-        
+
       - name: ⬇️ Checkout repo
         uses: actions/checkout@v3
         with:
@@ -37,11 +39,25 @@ jobs:
             exit 1
           fi
           echo "::set-output name=version::${IMAGE_TAG}"
+
       - name: 🔢 Get the commit hash
         id: get_commit
         run: |
           echo ::set-output name=sha_short::$(echo ${{ github.sha }} | cut -c1-7)
 
+      - name: 📛 Set the tags
+        id: set_tags
+        run: |
+          ref_without_tag=ghcr.io/triggerdotdev/trigger.dev
+          image_tags=$ref_without_tag:${{ steps.get_version.outputs.version }}
+
+          # if it's a versioned tag, also tag it as latest
+          if [[ "${{ github.ref_name }}" == v.docker.* ]]; then
+            image_tags=$image_tags,$ref_without_tag:latest
+          fi
+          
+          echo "IMAGE_TAGS=${image_tags}" >> "$GITHUB_OUTPUT"
+
       - name: 🐙 Login to GitHub Container Registry
         uses: docker/login-action@v2
         with:
@@ -54,6 +70,5 @@ jobs:
         with:
           file: ./docker/Dockerfile
           platforms: linux/amd64,linux/arm64
-          tags: |
-            ghcr.io/triggerdotdev/trigger.dev:${{ steps.get_version.outputs.version }}
+          tags: ${{ steps.set_tags.outputs.IMAGE_TAGS }}
           push: true

.github/workflows/publish-infra.yml

@@ -1,6 +1,7 @@
 name: "🚢 Publish Infra Images"
 
 on:
+  workflow_call:
   push:
     tags:
       - "infra-dev-*"
@@ -29,9 +30,6 @@ permissions:
   packages: write
   contents: read
 
-concurrency:
-  group: ${{ github.workflow }}-${{ github.ref }}
-
 env:
   AWS_REGION: us-east-1
 
@@ -39,7 +37,7 @@ jobs:
   build:
     strategy:
       matrix:
-        package: [coordinator, kubernetes-provider]
+        package: [coordinator, docker-provider, kubernetes-provider]
     runs-on: buildjet-16vcpu-ubuntu-2204
     env:
       DOCKER_BUILDKIT: "1"
@@ -48,20 +46,40 @@ jobs:
 
       - name: Generate image reference
         id: prep
-        # WARNING: This step expects the workflow to have been triggered by a specific tag format of: infra-${env}-*
         run: |
-          env=$(echo ${{ github.ref_name }} | cut -d- -f2)
-          sha=${GITHUB_SHA::7}
-          ts=$(date +%s)
+          # set image repo
           if [[ "${{ matrix.package }}" == *-provider ]]; then
-            provider_type=$(echo ${{ matrix.package }} | cut -d- -f1)
+            provider_type=$(echo "${{ matrix.package }}" | cut -d- -f1)
             repository=provider/${provider_type}
           else
-            repository=${{ matrix.package }}
+            repository="${{ matrix.package }}"
           fi
-          echo "IMAGE_TAG=${env}-${sha}-${ts}" >> "$GITHUB_OUTPUT"
           echo "REPOSITORY=${repository}" >> "$GITHUB_OUTPUT"
 
+          # set image tag
+          if [[ "${{ github.ref_type }}" == "tag" ]]; then
+            if [[ "${{ github.ref_name }}" == infra-*-* ]]; then
+              env=$(echo ${{ github.ref_name }} | cut -d- -f2)
+              sha=$(echo ${{ github.sha }} | head -c7)
+              ts=$(date +%s)
+              image_tag=${env}-${sha}-${ts}
+            elif [[ "${{ github.ref_name }}" == v.docker.* ]]; then
+              version="${GITHUB_REF_NAME#v.docker.}"
+              image_tag="v${version}"
+            elif [[ "${{ github.ref_name }}" == build-* ]]; then
+              image_tag="${GITHUB_REF_NAME#build-}"
+            else
+              echo "Invalid tag: ${{ github.ref_name }}"
+              exit 1
+            fi
+          elif [[ "${{ github.ref_name }}" == "main" ]]; then
+            image_tag="main"
+          else
+            echo "Invalid reference: ${{ github.ref }}"
+            exit 1
+          fi
+          echo "IMAGE_TAG=${image_tag}" >> "$GITHUB_OUTPUT"
+
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3
 
@@ -92,3 +110,12 @@ jobs:
           REGISTRY: ghcr.io/triggerdotdev
           REPOSITORY: ${{ steps.prep.outputs.REPOSITORY }}
           IMAGE_TAG: ${{ steps.prep.outputs.IMAGE_TAG }}
+
+      - name: 🐙 Push 'latest' to GitHub Container Registry
+        if: startsWith(github.ref_name, 'v.docker.')
+        run: |
+          docker tag infra_image $REGISTRY/$REPOSITORY:latest
+          docker push $REGISTRY/$REPOSITORY:latest
+        env:
+          REGISTRY: ghcr.io/triggerdotdev
+          REPOSITORY: ${{ steps.prep.outputs.REPOSITORY }}

.github/workflows/publish.yml

@@ -49,11 +49,18 @@ jobs:
     uses: ./.github/workflows/unit-tests.yml
     secrets: inherit
 
-  e2e:
-    uses: ./.github/workflows/e2e.yml
-    secrets: inherit
+  # e2e:
+  #   uses: ./.github/workflows/e2e.yml
+  #   with:
+  #     package: cli-v3
+  #   secrets: inherit
 
   publish:
-    needs: [typecheck, units, e2e]
+    needs: [typecheck, units]
     uses: ./.github/workflows/publish-docker.yml
     secrets: inherit
+
+  publish-infra:
+    needs: [typecheck, units]
+    uses: ./.github/workflows/publish-infra.yml
+    secrets: inherit

.vscode/launch.json

@@ -45,6 +45,14 @@
       "cwd": "${workspaceFolder}/references/v3-catalog",
       "sourceMaps": true
     },
+    {
+      "type": "node-terminal",
+      "request": "launch",
+      "name": "Debug V3 Management",
+      "command": "pnpm run management",
+      "cwd": "${workspaceFolder}/references/v3-catalog",
+      "sourceMaps": true
+    },
     {
       "type": "node",
       "request": "attach",

CONTRIBUTING.md

@@ -10,13 +10,14 @@ Thank you for helping us make Trigger.dev even better! 🤩
 
 The development branch is `main`. This is the branch that all pull
 requests should be made against. The changes on the `main`
-branch are tagged into a release monthly.
+branch are tagged into a release periodically.
 
 ### Prerequisites
 
-- [Node.js](https://nodejs.org/en) version >=16.x
-- [pnpm package manager](https://pnpm.io/installation) version 7
+- [Node.js](https://nodejs.org/en) version 20.11.1
+- [pnpm package manager](https://pnpm.io/installation) version 8.15.5
 - [Docker](https://www.docker.com/get-started/)
+- [protobuf](https://github.com/protocolbuffers/protobuf)
 
 ### Setup
 
@@ -33,15 +34,19 @@ branch are tagged into a release monthly.
    ```
    cd trigger.dev
    ```
-3. Install the required packages using pnpm.
+3. Ensure you are on the correct version of Node.js (20.11.1). If you are using `nvm`, there is an `.nvmrc` file that will automatically select the correct version of Node.js when you navigate to the repository.
+
+4. Run `corepack enable` to use the correct version of pnpm (`8.15.5`) as specified in the root `package.json` file.
+
+5. Install the required packages using pnpm.
    ```
    pnpm i
    ```
-4. Create your `.env` file
+6. Create your `.env` file
    ```
    cp .env.example .env
    ```
-5. Open it and generate a new value for `ENCRYPTION_KEY`:
+7. Open it and generate a new value for `ENCRYPTION_KEY`:
 
    `ENCRYPTION_KEY` is used to two-way encrypt OAuth access tokens and so you'll probably want to actually generate a unique value, and it must be a random 16 byte hex string. You can generate one with the following command:
 
@@ -51,7 +56,7 @@ branch are tagged into a release monthly.
 
    Feel free to update `SESSION_SECRET` and `MAGIC_LINK_SECRET` as well using the same method.
 
-6. Start Docker. This starts the required services like Postgres. If this is your first time using Docker, consider going through this [guide](DOCKER_INSTALLATION.md)
+8. Start Docker. This starts the required services like Postgres & Redis. If this is your first time using Docker, consider going through this [guide](DOCKER_INSTALLATION.md)
 
    ```
    pnpm run docker
@@ -59,19 +64,15 @@ branch are tagged into a release monthly.
 
    This will also start and run a local instance of [pgAdmin](https://www.pgadmin.org/) on [localhost:5480](http://localhost:5480), preconfigured with email `admin@example.com` and pwd `admin`. Then use `postgres` as the password to the Trigger.dev server.
 
-7. Migrate the database
+9. Migrate the database
    ```
    pnpm run db:migrate
    ```
-8. Build the app
-   ```
-   pnpm run build --filter webapp
-   ```
-9. Run the seed script
-   ```
-   pnpm run db:seed
-   ```
-10. Run the app. See the section below.
+10. Build the server app
+    ```
+    pnpm run build --filter webapp
+    ```
+11. Run the app. See the section below.
 
 ## Running
 
@@ -83,99 +84,96 @@ branch are tagged into a release monthly.
 
    It should run on port `3030`: [http://localhost:3030](http://localhost:3030/)
 
-2. Once the app is running click the magic link button and enter your email.
-3. Check your terminal, the magic link email should have printed out as following:
+2. Once the app is running click the magic link button and enter your email. You will automatically be logged in, since you are running locally. Create an Org and your first project in the dashboard.
 
-   ```sh
-   webapp:dev: Log in to Trigger.dev
-   webapp:dev:
-   webapp:dev: Click here to log in with this magic link
-   webapp:dev: [http://localhost:3030/magic?token=U2FsdGVkX18OvB0JxgaswTLCSbaRz%2FY82TN0EZWhSzFyZYwgG%2BIzKVTkeiaOtWfotPw7F8RwFzCHh53aBpMEu%2B%2B%2FItb%2FcJYh89MSjc3Pz92bevoEjqxSQ%2Ff%2BZbks09JOpqlBbYC3FzGWC8vuSVFBlxqLXxteSDLthZSUaC%2BS2LaA%2BJgp%2BLO7hgjAaC2lXbCHrM7MTgTdXOFt7i0Dvvuwz6%2BWY25RnfomZOPqDsyH0xz8Q2rzPTz0Xu53WSXrZ1hd]
-   webapp:dev:
-   webapp:dev: If you didn't try to log in, you can safely ignore this email.
-   ```
+## Manual testing using v3-catalog
 
-   Paste the magic link shown in your terminal into your browser to login.
+We use the `<root>/references/v3-catalog` subdirectory as a staging ground for testing changes to the SDK (`@trigger.dev/sdk` at `<root>/packages/trigger-sdk`), the Core package (`@trigger.dev/core` at `<root>packages/core`), the CLI (`trigger.dev` at `<root>/packages/cli-v3`) and the platform (The remix app at `<root>/apps/webapp`). The instructions below will get you started on using the `v3-catalog` for local development of Trigger.dev (v3).
 
-## Adding and running migrations
+### First-time setup
 
-1. Modify packages/database/prisma/schema.prisma file
-2. Change directory to the packages/database folder
-   ```sh
-   cd packages/database
-   ```
-3. Generate the Prisma client
+First, make sure you are running the webapp according to the instructions above. Then:
 
-   ```sh
-   pnpm run generate
-   ```
+1. In Postgres go to the "Organizations" table and on your org set the `v3Enabled` column to `true`.
 
-   The above updates the prisma client generated into node_modules/.prisma/client folder. This helps with typing of relevant prisma models. It ensures typescript
-   recognizes fields added or removed from a model and type-checks appropriately.
+2. Visit http://localhost:3030 in your browser and create a new V3 project called "v3-catalog". If you don't see an option for V3, you haven't set the `v3Enabled` flag to true.
 
-4. Create and apply the migrations
+3. In Postgres go to the "Projects" table and for the project you create change the `externalRef` to `yubjwjsfkxnylobaqvqz`.
 
-   ```
-   pnpm run db:migrate:dev
-   ```
-
-   This creates a migration file and executes the migrations against your database and applies changes to the database schema(s)
-
-5. Commit generated migrations as well as changes to the schema.prisma file
-6. If you're using VSCode you may need to restart the Typescript server in the webapp to get updated type inference. Open a TypeScript file, then open the Command Palette (View > Command Palette) and run `TypeScript: Restart TS server`.
-
-## Testing CLI changes
-
-To test CLI changes, follow the steps below:
-
-1. Build the CLI and watch for changes
+4. Build the CLI
 
 ```sh
-cd packages/cli
-pnpm run dev
-```
-
-2. Open a new Terminal window and run the webapp locally and then create a new project in the dashboard. Copy out the dev API key.
-
-3. Create a new temporary Next.js app in references directory
-
-```sh
-cd ./references
-pnpm create next-app@latest test-cli --ts --no-eslint --tailwind --app --src-dir --import-alias "@/*"
-```
-
-4. Then once that's finished, add the `@trigger.dev/cli` to the `devDependencies` of the newly created Next.js app's `package.json` file, like so:
-
-```json
-{
-  // other package.json properties
-  "devDependencies": { "@trigger.dev/cli": "workspace:*" }
-}
-```
-
-5. Back in the terminal, navigate into the reference, and initialize the CLI. When prompted, select `self-hosted` and enter `localhost:3030` if you are testing against the local instance of Trigger.dev, or you can just use the Trigger.dev cloud. When asked for an API key, use the key you copied earlier.
-
-```sh
-cd ./test-cli
+# Build the CLI
+pnpm run build --filter trigger.dev
+# Make it accessible to `pnpm exec`
 pnpm i
-pnpm exec trigger-cli init
 ```
 
-6. If you are just testing the `init` command, you can stop here. If you'd like to test the `dev` command, first start the Next.js app on port 3000:
+5. Change into the `<root>/references/v3-catalog` directory and authorize the CLI to the local server:
 
 ```sh
-pnpm run dev
+cd references/v3-catalog
+cp .env.example .env
+pnpm exec triggerdev login -a http://localhost:3030
 ```
 
-7. Open a new terminal window, and then run the `dev` command like so:
+This will open a new browser window and authorize the CLI against your local user account.
+
+You can optionally pass a `--profile` flag to the `login` command, which will allow you to use the CLI with separate accounts/servers. We suggest using a profile called `local` for your local development:
 
 ```sh
-pnpm exec trigger-cli dev
+cd references/v3-catalog
+pnpm exec triggerdev login -a http://localhost:3030 --profile local
+# later when you run the dev or deploy command:
+pnpm exec triggerdev dev --profile local
+pnpm exec triggerdev deploy --profile local
 ```
 
-8. Please remember to delete the temporary project you created after you've tested the changes, and before you raise a PR.
+### Running
 
-## Running end-to-end webapp tests
+The following steps should be followed any time you start working on a new feature you want to test in v3:
+
+1. Make sure the webapp is running on localhost:3030
+
+2. Open a terminal window and build the CLI and watch for changes
+
+```sh
+pnpm run dev --filter trigger.dev
+```
+
+2. Open a new terminal window, and anytime changes are made to the `@trigger.dev/core` package, you'll need to manually rebuild the CLI:
+
+```sh
+pnpm run build --filter trigger.dev
+```
+
+Note: You do not need to do the same for `@trigger.dev/sdk`, just core.
+
+3. Open another terminal window, and change into the `<root>/references/v3-catalog` directory.
+
+4. Run the `dev` command, which will register all the local tasks with the platform and allow you to start testing task execution:
+
+```sh
+# in <root>/references/v3-catalog
+pnpm exec triggerdev dev
+```
+
+If you want additional debug logging, you can use the `--log-level debug` flag:
+
+```sh
+# in <root>/references/v3-catalog
+pnpm exec triggerdev dev --log-level debug
+```
+
+5. If you make any changes in the CLI/Core/SDK, you'll need to `CTRL+C` to exit the `dev` command and restart it to pickup changes. Any changes to the files inside of the `v3-catalog/src/trigger` dir will automatically be rebuilt by the `dev` command.
+
+6. Navigate to the `v3-catalog` project in your local dashboard at localhost:3030 and you should see the list of tasks.
+
+7. Go to the "Test" page in the sidebar and select a task. Then enter a payload and click "Run test". You can tell what the payloads should be by looking at the relevant task file inside the `/references/v3-catalog/src/trigger` folder. Many of them accept an empty payload.
+
+8. Feel free to add additional files in `v3-catalog/src/trigger` to test out specific aspects of the system, or add in edge cases.
+
+## Running end-to-end webapp tests (deprecated)
 
 To run the end-to-end tests, follow the steps below:
 
@@ -223,6 +221,26 @@ The end-to-end tests use a `setup` and `teardown` script to seed the database wi
 pnpm run db:studio
 ```
 
+## Adding and running migrations
+
+1. Modify packages/database/prisma/schema.prisma file
+2. Change directory to the packages/database folder
+
+   ```sh
+   cd packages/database
+   ```
+
+3. Create and apply the migrations
+
+   ```
+   pnpm run db:migrate:dev
+   ```
+
+   This creates a migration file and executes the migrations against your database and applies changes to the database schema(s)
+
+4. Commit generated migrations as well as changes to the schema.prisma file
+5. If you're using VSCode you may need to restart the Typescript server in the webapp to get updated type inference. Open a TypeScript file, then open the Command Palette (View > Command Palette) and run `TypeScript: Restart TS server`.
+
 ## Add sample jobs
 
 The [references/job-catalog](./references/job-catalog/) project defines simple jobs you can get started with.

README.md

@@ -5,7 +5,7 @@
   <img alt="Trigger.dev logo" src="https://imagedelivery.net/3TbraffuDZ4aEf8KWOmI_w/a45d1fa2-0ae8-4a39-4409-f4f934bfae00/public">
 </picture>
   
-### The open source background jobs framework
+### The open source background jobs platform
 
 [Discord](https://trigger.dev/discord) | [Website](https://trigger.dev) | [Issues](https://github.com/triggerdotdev/trigger.dev/issues) | [Docs](https://trigger.dev/docs)
 
@@ -14,7 +14,9 @@
 
 </div>
 
-# About Trigger.dev
+> The Trigger.dev v3 developer preview is now open. For more information and to get early access, check out our [developer preview launch post](https://trigger.dev/blog/v3-developer-preview-launch/).
+
+## About Trigger.dev
 
 Create long-running jobs directly in your codebase with features like API integrations, webhooks, scheduling and delays.
 
@@ -48,30 +50,6 @@ View every Task in every Run so you can tell exactly what happened.
 
 Easily integrate with hundreds of third-party APIs – including your own. Use API keys (which never leave your server) or let us handle OAuth for you. Install our integration packages and easily subscribe to webhooks and perform common tasks, or you can easily use your existing favorite Node.JS SDKs and get resumability and idempotency through our `runTask` function.
 
-## Our progress
-
-We’re building the most comprehensive and easy-to-use background jobs framework for developers.
-
-Click the links to join the discussions about our upcoming features.
-
-| Feature                                                                              | What it does                                        | Status |
-| ------------------------------------------------------------------------------------ | --------------------------------------------------- | ------ |
-| Integration kit                                                                      | Official Trigger.dev integrations or build your own | ✅     |
-| Self-hosting                                                                         | Host the platform yourself                          | ✅     |
-| Cloud                                                                                | Just write code, no deployment required             | ✅     |
-| Dashboard                                                                            | View every Task in every Run                        | ✅     |
-| Serverless                                                                           | Long-running Jobs on your serverless backend        | ✅     |
-| React hooks                                                                          | Easily update your UI with Job progress             | ✅     |
-| React frameworks                                                                     | Support for Remix, Astro, RedwoodJS & more          | ✅     |
-| [Background tasks](https://github.com/triggerdotdev/trigger.dev/discussions/400)     | Offload long or intense Tasks to our infrastructure | 🛠️     |
-| [Long-running servers](https://github.com/triggerdotdev/trigger.dev/discussions/430) | Run Jobs on your long-running backend               | 🛠️     |
-| Polling Triggers                                                                     | Subscribe to changes without webhooks               | 🕝     |
-| Vercel integration                                                                   | Easy deploy and preview environment support         | 🕝     |
-| Streaming                                                                            | Receive data from your Jobs in realtime             | 🕝     |
-| 100+ integrations                                                                    | Comprehensive support for popular APIs              | 🕝     |
-| [Trigger.dev Connect](https://github.com/triggerdotdev/trigger.dev/discussions/441)  | Use integrations signed in as your users            | 🕝     |
-| File IO                                                                              | Create Tasks that have file outputs                 | 🕝     |
-
 # Getting started
 
 Visit our docs [here](https://trigger.dev/docs).

apps/coordinator/Containerfile

@@ -1,19 +1,19 @@
 # syntax=docker/dockerfile:labs
 
-FROM node:18-bullseye-slim@sha256:a4edd54dcfdcacc8a4100fee71498e8671d99556a1acf5614539214a70092426 AS node-18
+FROM node:20-bookworm-slim@sha256:72f2f046a5f8468db28730b990b37de63ce93fd1a72a40f531d6aa82afdf0d46 AS node-20
 
 WORKDIR /app
 
-FROM node-18 AS pruner
+FROM node-20 AS pruner
 
 COPY --chown=node:node . .
 RUN npx -q turbo@1.10.9 prune --scope=coordinator --docker
 RUN find . -name "node_modules" -type d -prune -exec rm -rf '{}' +
 
-FROM node-18 AS base
+FROM node-20 AS base
 
 RUN apt-get update \
-  && apt-get install -y buildah ca-certificates dumb-init \
+  && apt-get install -y buildah ca-certificates dumb-init docker.io \
   && rm -rf /var/lib/apt/lists/*
 
 COPY --chown=node:node .gitignore .gitignore

apps/coordinator/package.json

@@ -21,8 +21,8 @@
     "execa": "^8.0.1",
     "nanoid": "^5.0.6",
     "prom-client": "^15.1.0",
-    "socket.io": "^4.7.4",
-    "socket.io-client": "^4.7.4"
+    "socket.io": "4.7.4",
+    "socket.io-client": "4.7.4"
   },
   "devDependencies": {
     "@types/node": "^18",

apps/coordinator/src/backoff.ts

@@ -0,0 +1,247 @@
+type ExponentialBackoffType = "NoJitter" | "FullJitter" | "EqualJitter";
+
+type ExponentialBackoffOptions = {
+  base: number;
+  factor: number;
+  min: number;
+  max: number;
+  maxRetries: number;
+  maxElapsed: number;
+};
+
+class StopRetrying extends Error {
+  constructor(message?: string) {
+    super(message);
+    this.name = "StopRetrying";
+  }
+}
+
+export class ExponentialBackoff {
+  #retries: number = 0;
+
+  #type: ExponentialBackoffType;
+  #base: number;
+  #factor: number;
+
+  #min: number;
+  #max: number;
+
+  #maxRetries: number;
+  #maxElapsed: number;
+
+  constructor(type?: ExponentialBackoffType, opts: Partial<ExponentialBackoffOptions> = {}) {
+    this.#type = type ?? "NoJitter";
+    this.#base = opts.base ?? 2;
+    this.#factor = opts.factor ?? 1;
+
+    this.#min = opts.min ?? -Infinity;
+    this.#max = opts.max ?? Infinity;
+
+    this.#maxRetries = opts.maxRetries ?? Infinity;
+    this.#maxElapsed = opts.maxElapsed ?? Infinity;
+  }
+
+  #clone() {
+    return new ExponentialBackoff(this.#type, {
+      base: this.#base,
+      factor: this.#factor,
+      min: this.#min,
+      max: this.#max,
+      maxRetries: this.#maxRetries,
+      maxElapsed: this.#maxElapsed,
+    });
+  }
+
+  type(type?: ExponentialBackoffType) {
+    if (typeof type !== "undefined") {
+      this.#type = type;
+    }
+    return this.#clone();
+  }
+
+  base(base?: number) {
+    if (typeof base !== "undefined") {
+      this.#base = base;
+    }
+    return this.#clone();
+  }
+
+  factor(factor?: number) {
+    if (typeof factor !== "undefined") {
+      this.#factor = factor;
+    }
+    return this.#clone();
+  }
+
+  min(min?: number) {
+    if (typeof min !== "undefined") {
+      this.#min = min;
+    }
+    return this.#clone();
+  }
+
+  max(max?: number) {
+    if (typeof max !== "undefined") {
+      this.#max = max;
+    }
+    return this.#clone();
+  }
+
+  maxRetries(maxRetries?: number) {
+    if (typeof maxRetries !== "undefined") {
+      this.#maxRetries = maxRetries;
+    }
+    return this.#clone();
+  }
+
+  maxElapsed(maxElapsed?: number) {
+    if (typeof maxElapsed !== "undefined") {
+      this.#maxElapsed = maxElapsed;
+    }
+    return this.#clone();
+  }
+
+  retries(retries?: number) {
+    if (typeof retries !== "undefined") {
+      if (retries > this.#maxRetries) {
+        console.error(
+          `Can't set retries ${retries} higher than maxRetries (${
+            this.#maxRetries
+          }), setting to maxRetries instead.`
+        );
+        this.#retries = this.#maxRetries;
+      } else {
+        this.#retries = retries;
+      }
+    }
+    return this.#clone();
+  }
+
+  async *retryAsync(maxRetries: number = this.#maxRetries ?? Infinity) {
+    let elapsed = 0;
+    let retry = 0;
+
+    while (retry <= maxRetries) {
+      const delay = this.delay(retry);
+      elapsed += delay;
+
+      if (elapsed > this.#maxElapsed) {
+        break;
+      }
+
+      yield {
+        delay: {
+          seconds: delay,
+          milliseconds: delay * 1000,
+        },
+        retry,
+      };
+
+      retry++;
+    }
+  }
+
+  async *[Symbol.asyncIterator]() {
+    yield* this.retryAsync();
+  }
+
+  delay(retries: number = this.#retries, jitter: boolean = true) {
+    if (retries > this.#maxRetries) {
+      console.error(
+        `Can't set retries ${retries} higher than maxRetries (${
+          this.#maxRetries
+        }), setting to maxRetries instead.`
+      );
+      retries = this.#maxRetries;
+    }
+
+    let delay = this.#factor * this.#base ** retries;
+
+    switch (this.#type) {
+      case "NoJitter": {
+        break;
+      }
+      case "FullJitter": {
+        if (!jitter) {
+          delay = 0;
+          break;
+        }
+
+        delay *= Math.random();
+        break;
+      }
+      case "EqualJitter": {
+        if (!jitter) {
+          delay *= 0.5;
+          break;
+        }
+
+        delay *= 0.5 * (1 + Math.random());
+        break;
+      }
+      default: {
+        throw new Error(`Unknown backoff type: ${this.#type}`);
+      }
+    }
+
+    delay = Math.min(delay, this.#max);
+    delay = Math.max(delay, this.#min);
+    delay = Math.round(delay);
+
+    return delay;
+  }
+
+  elapsed(retries: number = this.#retries, jitter: boolean = true) {
+    let elapsed = 0;
+
+    for (let i = 0; i <= retries; i++) {
+      elapsed += this.delay(i, jitter);
+    }
+
+    const total = elapsed;
+
+    let days = 0;
+    if (elapsed > 3600 * 24) {
+      days = Math.floor(elapsed / 3600 / 24);
+      elapsed -= days * 3600 * 24;
+    }
+
+    let hours = 0;
+    if (elapsed > 3600) {
+      hours = Math.floor(elapsed / 3600);
+      elapsed -= hours * 3600;
+    }
+
+    let minutes = 0;
+    if (elapsed > 60) {
+      minutes = Math.floor(elapsed / 60);
+      elapsed -= minutes * 60;
+    }
+
+    const seconds = elapsed;
+
+    return {
+      seconds,
+      minutes,
+      hours,
+      days,
+      total,
+    };
+  }
+
+  reset() {
+    this.#retries = 0;
+    return this;
+  }
+
+  next() {
+    this.#retries++;
+    return this.delay();
+  }
+
+  stop() {
+    throw new StopRetrying();
+  }
+
+  static StopRetrying = StopRetrying;
+}

apps/docker-provider/.env.example

@@ -4,6 +4,8 @@ PLATFORM_WS_PORT=3030
 PLATFORM_SECRET=provider-secret
 SECURE_CONNECTION=false
 
+OTEL_EXPORTER_OTLP_ENDPOINT=http://0.0.0.0:3030/otel
+
 # Use this if you are on macOS
 # COORDINATOR_HOST="host.docker.internal"
 # OTEL_EXPORTER_OTLP_ENDPOINT="http://host.docker.internal:4318"

apps/docker-provider/Containerfile

@@ -1,16 +1,47 @@
-# syntax=docker/dockerfile:labs
-
-FROM node:18-slim AS base
-
-RUN apt-get update \
-  && apt-get install -y dumb-init
-
-FROM base
+FROM node:20-alpine@sha256:7a91aa397f2e2dfbfcdad2e2d72599f374e0b0172be1d86eeb73f1d33f36a4b2 AS node-20-alpine
 
 WORKDIR /app
 
-COPY --chown=node dist/index.mjs /app/
+FROM node-20-alpine AS pruner
+
+COPY --chown=node:node . .
+RUN npx -q turbo@1.10.9 prune --scope=docker-provider --docker
+RUN find . -name "node_modules" -type d -prune -exec rm -rf '{}' +
+
+FROM node-20-alpine AS base
+
+RUN apk add --no-cache dumb-init docker
+
+COPY --chown=node:node .gitignore .gitignore
+COPY --from=pruner --chown=node:node /app/out/json/ .
+COPY --from=pruner --chown=node:node /app/out/pnpm-lock.yaml ./pnpm-lock.yaml
+COPY --from=pruner --chown=node:node /app/out/pnpm-workspace.yaml ./pnpm-workspace.yaml
+
+FROM base AS dev-deps
+RUN corepack enable
+ENV NODE_ENV development
+
+RUN --mount=type=cache,id=pnpm,target=/root/.local/share/pnpm/store pnpm fetch --no-frozen-lockfile
+RUN --mount=type=cache,id=pnpm,target=/root/.local/share/pnpm/store pnpm install --ignore-scripts --no-frozen-lockfile
+
+FROM base AS builder
+RUN corepack enable
+
+COPY --from=pruner --chown=node:node /app/out/full/ .
+COPY --from=dev-deps --chown=node:node /app/ .
+COPY --chown=node:node turbo.json turbo.json
+
+RUN pnpm run -r --filter docker-provider build:bundle
+
+FROM base AS runner
+
+RUN corepack enable
+ENV NODE_ENV production
+
+COPY --from=builder --chown=node:node /app/apps/docker-provider/dist/index.mjs ./index.mjs
 
 EXPOSE 8000
 
-ENTRYPOINT [ "/usr/bin/dumb-init", "--", "/usr/local/bin/node", "/app/index.mjs" ]
+USER node
+
+CMD [ "/usr/bin/dumb-init", "--", "/usr/local/bin/node", "./index.mjs" ]

apps/docker-provider/src/index.ts

@@ -6,6 +6,8 @@ import {
   TaskOperationsRestoreOptions,
   TaskOperationsCreateOptions,
   TaskOperationsIndexOptions,
+  isExecaChildProcess,
+  testDockerCheckpoint,
 } from "@trigger.dev/core-apps";
 import { setTimeout } from "node:timers/promises";
 import { PostStartCauses, PreStopCauses } from "@trigger.dev/core/v3";
@@ -13,75 +15,68 @@ import { PostStartCauses, PreStopCauses } from "@trigger.dev/core/v3";
 const MACHINE_NAME = process.env.MACHINE_NAME || "local";
 const COORDINATOR_PORT = process.env.COORDINATOR_PORT || 8020;
 const COORDINATOR_HOST = process.env.COORDINATOR_HOST || "127.0.0.1";
+
 const OTEL_EXPORTER_OTLP_ENDPOINT =
   process.env.OTEL_EXPORTER_OTLP_ENDPOINT || "http://0.0.0.0:4318";
 
+const FORCE_CHECKPOINT_SIMULATION = ["1", "true"].includes(
+  process.env.FORCE_CHECKPOINT_SIMULATION ?? "true"
+);
+
 const logger = new SimpleLogger(`[${MACHINE_NAME}]`);
 
-type InitializeReturn = {
+type TaskOperationsInitReturn = {
   canCheckpoint: boolean;
   willSimulate: boolean;
 };
 
-function isExecaChildProcess(maybeExeca: unknown): maybeExeca is Awaited<ExecaChildProcess> {
-  return typeof maybeExeca === "object" && maybeExeca !== null && "escapedCommand" in maybeExeca;
-}
-
 class DockerTaskOperations implements TaskOperations {
   #initialized = false;
   #canCheckpoint = false;
 
   constructor(private opts = { forceSimulate: false }) {}
 
-  async #initialize(): Promise<InitializeReturn> {
+  async init(): Promise<TaskOperationsInitReturn> {
     if (this.#initialized) {
-      return this.#getInitializeReturn();
+      return this.#getInitReturn(this.#canCheckpoint);
     }
 
     logger.log("Initializing task operations");
 
-    if (this.opts.forceSimulate) {
-      logger.log("Forced simulation enabled. Will simulate regardless of checkpoint support.");
+    const testCheckpoint = await testDockerCheckpoint();
+
+    if (testCheckpoint.ok) {
+      return this.#getInitReturn(true);
     }
 
-    try {
-      await $`criu --version`;
-    } catch (error) {
-      logger.error("No checkpoint support: Missing CRIU binary. Will simulate instead.");
-      this.#canCheckpoint = false;
-      this.#initialized = true;
-
-      return this.#getInitializeReturn();
-    }
-
-    try {
-      await $`docker checkpoint`;
-    } catch (error) {
-      logger.error("No checkpoint support: Docker needs to have experimental features enabled");
-      logger.error("Will simulate instead");
-      this.#canCheckpoint = false;
-      this.#initialized = true;
-
-      return this.#getInitializeReturn();
-    }
-
-    logger.log("Full checkpoint support!");
-
-    this.#initialized = true;
-    this.#canCheckpoint = true;
-
-    return this.#getInitializeReturn();
+    logger.error(testCheckpoint.message, testCheckpoint.error);
+    return this.#getInitReturn(false);
   }
 
-  #getInitializeReturn(): InitializeReturn {
+  #getInitReturn(canCheckpoint: boolean): TaskOperationsInitReturn {
+    this.#initialized = true;
+    this.#canCheckpoint = canCheckpoint;
+
+    if (canCheckpoint) {
+      logger.log("Full checkpoint support!");
+    }
+
+    const willSimulate = !canCheckpoint || this.opts.forceSimulate;
+
+    if (willSimulate) {
+      logger.log("Simulation mode enabled. Containers will be paused, not checkpointed.", {
+        forceSimulate: this.opts.forceSimulate,
+      });
+    }
+
     return {
-      canCheckpoint: this.#canCheckpoint,
-      willSimulate: !this.#canCheckpoint || this.opts.forceSimulate,
+      canCheckpoint,
+      willSimulate,
     };
   }
 
   async index(opts: TaskOperationsIndexOptions) {
-    await this.#initialize();
+    await this.init();
 
     const containerName = this.#getIndexContainerName(opts.shortCode);
 
@@ -90,60 +85,51 @@ class DockerTaskOperations implements TaskOperations {
       port: COORDINATOR_PORT,
     });
 
-    try {
-      logger.debug(
-        await execa("docker", [
-          "run",
-          "--network=host",
-          "--rm",
-          `--env=INDEX_TASKS=true`,
-          `--env=TRIGGER_SECRET_KEY=${opts.apiKey}`,
-          `--env=TRIGGER_API_URL=${opts.apiUrl}`,
-          `--env=TRIGGER_ENV_ID=${opts.envId}`,
-          `--env=OTEL_EXPORTER_OTLP_ENDPOINT=${OTEL_EXPORTER_OTLP_ENDPOINT}`,
-          `--env=POD_NAME=${containerName}`,
-          `--env=COORDINATOR_HOST=${COORDINATOR_HOST}`,
-          `--env=COORDINATOR_PORT=${COORDINATOR_PORT}`,
-          `--name=${containerName}`,
-          `${opts.imageRef}`,
-        ])
-      );
-    } catch (error: any) {
-      if (!isExecaChildProcess(error)) {
-        throw error;
-      }
-
-      logger.error("Index failed:", {
-        opts,
-        exitCode: error.exitCode,
-        escapedCommand: error.escapedCommand,
-        stdout: error.stdout,
-        stderr: error.stderr,
-      });
-    }
+    logger.debug(
+      await execa("docker", [
+        "run",
+        "--network=host",
+        "--rm",
+        `--env=INDEX_TASKS=true`,
+        `--env=TRIGGER_SECRET_KEY=${opts.apiKey}`,
+        `--env=TRIGGER_API_URL=${opts.apiUrl}`,
+        `--env=TRIGGER_ENV_ID=${opts.envId}`,
+        `--env=OTEL_EXPORTER_OTLP_ENDPOINT=${OTEL_EXPORTER_OTLP_ENDPOINT}`,
+        `--env=POD_NAME=${containerName}`,
+        `--env=COORDINATOR_HOST=${COORDINATOR_HOST}`,
+        `--env=COORDINATOR_PORT=${COORDINATOR_PORT}`,
+        `--name=${containerName}`,
+        `${opts.imageRef}`,
+      ])
+    );
   }
 
   async create(opts: TaskOperationsCreateOptions) {
-    await this.#initialize();
+    await this.init();
 
     const containerName = this.#getRunContainerName(opts.runId);
 
+    const runArgs = [
+      "run",
+      "--network=host",
+      "--detach",
+      `--env=TRIGGER_ENV_ID=${opts.envId}`,
+      `--env=TRIGGER_RUN_ID=${opts.runId}`,
+      `--env=OTEL_EXPORTER_OTLP_ENDPOINT=${OTEL_EXPORTER_OTLP_ENDPOINT}`,
+      `--env=POD_NAME=${containerName}`,
+      `--env=COORDINATOR_HOST=${COORDINATOR_HOST}`,
+      `--env=COORDINATOR_PORT=${COORDINATOR_PORT}`,
+      `--name=${containerName}`,
+    ];
+
+    if (process.env.ENFORCE_MACHINE_PRESETS) {
+      runArgs.push(`--cpus=${opts.machine.cpu}`, `--memory=${opts.machine.memory}G`);
+    }
+
+    runArgs.push(`${opts.image}`);
+
     try {
-      logger.debug(
-        await execa("docker", [
-          "run",
-          "--network=host",
-          "--detach",
-          `--env=TRIGGER_ENV_ID=${opts.envId}`,
-          `--env=TRIGGER_RUN_ID=${opts.runId}`,
-          `--env=OTEL_EXPORTER_OTLP_ENDPOINT=${OTEL_EXPORTER_OTLP_ENDPOINT}`,
-          `--env=POD_NAME=${containerName}`,
-          `--env=COORDINATOR_HOST=${COORDINATOR_HOST}`,
-          `--env=COORDINATOR_PORT=${COORDINATOR_PORT}`,
-          `--name=${containerName}`,
-          `${opts.image}`,
-        ])
-      );
+      logger.debug(await execa("docker", runArgs));
     } catch (error) {
       if (!isExecaChildProcess(error)) {
         throw error;
@@ -160,7 +146,7 @@ class DockerTaskOperations implements TaskOperations {
   }
 
   async restore(opts: TaskOperationsRestoreOptions) {
-    await this.#initialize();
+    await this.init();
 
     const containerName = this.#getRunContainerName(opts.runId);
 
@@ -189,7 +175,7 @@ class DockerTaskOperations implements TaskOperations {
   }
 
   async delete(opts: { runId: string }) {
-    await this.#initialize();
+    await this.init();
 
     const containerName = this.#getRunContainerName(opts.runId);
     await this.#sendPreStop(containerName);
@@ -198,7 +184,7 @@ class DockerTaskOperations implements TaskOperations {
   }
 
   async get(opts: { runId: string }) {
-    await this.#initialize();
+    await this.init();
 
     logger.log("noop: get");
   }
@@ -278,7 +264,7 @@ class DockerTaskOperations implements TaskOperations {
 }
 
 const provider = new ProviderShell({
-  tasks: new DockerTaskOperations({ forceSimulate: true }),
+  tasks: new DockerTaskOperations({ forceSimulate: FORCE_CHECKPOINT_SIMULATION }),
   type: "docker",
 });
 

apps/kubernetes-provider/Containerfile

@@ -1,14 +1,14 @@
-FROM node:18-alpine@sha256:ca9f6cb0466f9638e59e0c249d335a07c867cd50c429b5c7830dda1bed584649 AS node-18-alpine
+FROM node:20-alpine@sha256:7a91aa397f2e2dfbfcdad2e2d72599f374e0b0172be1d86eeb73f1d33f36a4b2 AS node-20-alpine
 
 WORKDIR /app
 
-FROM node-18-alpine AS pruner
+FROM node-20-alpine AS pruner
 
 COPY --chown=node:node . .
 RUN npx -q turbo@1.10.9 prune --scope=kubernetes-provider --docker
 RUN find . -name "node_modules" -type d -prune -exec rm -rf '{}' +
 
-FROM node-18-alpine AS base
+FROM node-20-alpine AS base
 
 RUN apk add --no-cache dumb-init
 

apps/kubernetes-provider/src/index.ts

@@ -7,14 +7,21 @@ import {
   TaskOperationsIndexOptions,
   TaskOperationsRestoreOptions,
 } from "@trigger.dev/core-apps";
-import { Machine, PostStartCauses, PreStopCauses, EnvironmentType } from "@trigger.dev/core/v3";
+import {
+  MachinePreset,
+  PostStartCauses,
+  PreStopCauses,
+  EnvironmentType,
+} from "@trigger.dev/core/v3";
 import { randomUUID } from "crypto";
 import { TaskMonitor } from "./taskMonitor";
+import { PodCleaner } from "./podCleaner";
 
 const RUNTIME_ENV = process.env.KUBERNETES_PORT ? "kubernetes" : "local";
 const NODE_NAME = process.env.NODE_NAME || "local";
 const OTEL_EXPORTER_OTLP_ENDPOINT =
   process.env.OTEL_EXPORTER_OTLP_ENDPOINT ?? "http://0.0.0.0:4318";
+const POD_CLEANER_INTERVAL_SECONDS = Number(process.env.POD_CLEANER_INTERVAL_SECONDS || "300");
 
 const logger = new SimpleLogger(`[${NODE_NAME}]`);
 logger.log(`running in ${RUNTIME_ENV} mode`);
@@ -46,6 +53,10 @@ class KubernetesTaskOperations implements TaskOperations {
     this.#k8sApi = this.#createK8sApi();
   }
 
+  async init() {
+    // noop
+  }
+
   async index(opts: TaskOperationsIndexOptions) {
     await this.#createJob(
       {
@@ -80,8 +91,8 @@ class KubernetesTaskOperations implements TaskOperations {
                   ],
                   resources: {
                     limits: {
-                      cpu: "250m",
-                      memory: "0.5G",
+                      cpu: "1",
+                      memory: "1G",
                       "ephemeral-storage": "2Gi",
                     },
                   },
@@ -211,7 +222,8 @@ class KubernetesTaskOperations implements TaskOperations {
             },
             {
               name: "populate-taskinfo",
-              image: "busybox",
+              image: "registry.digitalocean.com/trigger/busybox",
+              imagePullPolicy: "IfNotPresent",
               command: ["/bin/sh", "-c"],
               args: ["printenv COORDINATOR_HOST | tee /etc/taskinfo/coordinator-host"],
               env: [
@@ -314,6 +326,9 @@ class KubernetesTaskOperations implements TaskOperations {
         {
           name: "registry-trigger",
         },
+        {
+          name: "registry-trigger-failover",
+        },
       ],
       nodeSelector: {
         nodetype: "worker",
@@ -389,10 +404,10 @@ class KubernetesTaskOperations implements TaskOperations {
     };
   }
 
-  #getResourcesFromMachineConfig(config: Machine): ComputeResources {
+  #getResourcesFromMachineConfig(preset: MachinePreset): ComputeResources {
     return {
-      cpu: `${config.cpu}`,
-      memory: `${config.memory}G`,
+      cpu: `${preset.cpu}`,
+      memory: `${preset.memory}G`,
     };
   }
 
@@ -514,27 +529,28 @@ provider.listen();
 
 const taskMonitor = new TaskMonitor({
   runtimeEnv: RUNTIME_ENV,
-  onIndexFailure: async (deploymentId, failureInfo) => {
-    logger.log("Indexing failed", { deploymentId, failureInfo });
+  onIndexFailure: async (deploymentId, details) => {
+    logger.log("Indexing failed", { deploymentId, details });
 
     try {
       provider.platformSocket.send("INDEXING_FAILED", {
         deploymentId,
         error: {
-          name: `Crashed with exit code ${failureInfo.exitCode}`,
-          message: failureInfo.reason,
-          stack: failureInfo.logs,
+          name: `Crashed with exit code ${details.exitCode}`,
+          message: details.reason,
+          stack: details.logs,
         },
+        overrideCompletion: details.overrideCompletion,
       });
     } catch (error) {
       logger.error(error);
     }
   },
-  onRunFailure: async (runId, failureInfo) => {
-    logger.log("Run failed:", { runId, failureInfo });
+  onRunFailure: async (runId, details) => {
+    logger.log("Run failed:", { runId, details });
 
     try {
-      provider.platformSocket.send("WORKER_CRASHED", { runId, ...failureInfo });
+      provider.platformSocket.send("WORKER_CRASHED", { runId, ...details });
     } catch (error) {
       logger.error(error);
     }
@@ -542,3 +558,11 @@ const taskMonitor = new TaskMonitor({
 });
 
 taskMonitor.start();
+
+const podCleaner = new PodCleaner({
+  runtimeEnv: RUNTIME_ENV,
+  namespace: "default",
+  intervalInSeconds: POD_CLEANER_INTERVAL_SECONDS,
+});
+
+podCleaner.start();

apps/kubernetes-provider/src/podCleaner.ts

@@ -0,0 +1,264 @@
+import * as k8s from "@kubernetes/client-node";
+import { SimpleLogger } from "@trigger.dev/core-apps";
+
+type PodCleanerOptions = {
+  runtimeEnv: "local" | "kubernetes";
+  namespace?: string;
+  intervalInSeconds?: number;
+};
+
+export class PodCleaner {
+  private enabled = false;
+  private namespace = "default";
+  private intervalInSeconds = 300;
+
+  private logger = new SimpleLogger("[PodCleaner]");
+  private k8sClient: {
+    core: k8s.CoreV1Api;
+    kubeConfig: k8s.KubeConfig;
+  };
+
+  constructor(private opts: PodCleanerOptions) {
+    if (opts.namespace) {
+      this.namespace = opts.namespace;
+    }
+
+    if (opts.intervalInSeconds) {
+      this.intervalInSeconds = opts.intervalInSeconds;
+    }
+
+    this.k8sClient = this.#createK8sClient();
+  }
+
+  #createK8sClient() {
+    const kubeConfig = new k8s.KubeConfig();
+
+    if (this.opts.runtimeEnv === "local") {
+      kubeConfig.loadFromDefault();
+    } else if (this.opts.runtimeEnv === "kubernetes") {
+      kubeConfig.loadFromCluster();
+    } else {
+      throw new Error(`Unsupported runtime environment: ${this.opts.runtimeEnv}`);
+    }
+
+    return {
+      core: kubeConfig.makeApiClient(k8s.CoreV1Api),
+      kubeConfig: kubeConfig,
+    };
+  }
+
+  #isRecord(candidate: unknown): candidate is Record<string, unknown> {
+    if (typeof candidate !== "object" || candidate === null) {
+      return false;
+    } else {
+      return true;
+    }
+  }
+
+  #logK8sError(err: unknown, debugOnly = false) {
+    if (debugOnly) {
+      this.logger.debug("K8s API Error", err);
+    } else {
+      this.logger.error("K8s API Error", err);
+    }
+  }
+
+  #handleK8sError(err: unknown) {
+    if (!this.#isRecord(err) || !this.#isRecord(err.body)) {
+      this.#logK8sError(err);
+      return;
+    }
+
+    this.#logK8sError(err, true);
+
+    if (typeof err.body.message === "string") {
+      this.#logK8sError({ message: err.body.message });
+      return;
+    }
+
+    this.#logK8sError({ body: err.body });
+  }
+
+  async #deletePods(opts: {
+    namespace: string;
+    dryRun?: boolean;
+    fieldSelector?: string;
+    labelSelector?: string;
+  }) {
+    return await this.k8sClient.core
+      .deleteCollectionNamespacedPod(
+        opts.namespace,
+        undefined, // pretty
+        undefined, // continue
+        opts.dryRun ? "All" : undefined,
+        opts.fieldSelector,
+        undefined, // gracePeriodSeconds
+        opts.labelSelector
+      )
+      .catch(this.#handleK8sError.bind(this));
+  }
+
+  async #deleteCompletedRuns() {
+    this.logger.log("Deleting completed runs");
+
+    const start = Date.now();
+
+    const result = await this.#deletePods({
+      namespace: this.namespace,
+      fieldSelector: "status.phase=Succeeded",
+      labelSelector: "app=task-run",
+    });
+
+    const elapsedMs = Date.now() - start;
+
+    if (!result) {
+      this.logger.log("Deleting completed runs: No delete result", { elapsedMs });
+      return;
+    }
+
+    const total = (result.response as any)?.body?.items?.length ?? 0;
+
+    this.logger.log("Deleting completed runs: Done", { total, elapsedMs });
+  }
+
+  async #deleteFailedRuns() {
+    this.logger.log("Deleting failed runs");
+
+    const start = Date.now();
+
+    const result = await this.#deletePods({
+      namespace: this.namespace,
+      fieldSelector: "status.phase=Failed",
+      labelSelector: "app=task-run",
+    });
+
+    const elapsedMs = Date.now() - start;
+
+    if (!result) {
+      this.logger.log("Deleting failed runs: No delete result", { elapsedMs });
+      return;
+    }
+
+    const total = (result.response as any)?.body?.items?.length ?? 0;
+
+    this.logger.log("Deleting failed runs: Done", { total, elapsedMs });
+  }
+
+  async #deleteUnrecoverableRuns() {
+    await this.#deletePods({
+      namespace: this.namespace,
+      fieldSelector: "status.phase=?",
+      labelSelector: "app=task-run",
+    });
+  }
+
+  async start() {
+    this.enabled = true;
+    this.logger.log("Starting");
+
+    const completedInterval = setInterval(async () => {
+      if (!this.enabled) {
+        clearInterval(completedInterval);
+        return;
+      }
+
+      try {
+        await this.#deleteCompletedRuns();
+      } catch (error) {
+        this.logger.error("Error deleting completed runs", error);
+      }
+    }, this.intervalInSeconds * 1000);
+
+    const failedInterval = setInterval(
+      async () => {
+        if (!this.enabled) {
+          clearInterval(failedInterval);
+          return;
+        }
+
+        try {
+          await this.#deleteFailedRuns();
+        } catch (error) {
+          this.logger.error("Error deleting completed runs", error);
+        }
+      },
+      // Use a longer interval for failed runs. This is only a backup in case the task monitor fails.
+      2 * this.intervalInSeconds * 1000
+    );
+
+    // this.#launchTests();
+  }
+
+  async stop() {
+    if (!this.enabled) {
+      return;
+    }
+
+    this.enabled = false;
+    this.logger.log("Shutting down..");
+  }
+
+  async #launchTests() {
+    const createPod = async (
+      container: k8s.V1Container,
+      name: string,
+      labels?: Record<string, string>
+    ) => {
+      this.logger.log("Creating pod:", name);
+
+      const pod = {
+        metadata: {
+          name,
+          labels,
+        },
+        spec: {
+          restartPolicy: "Never",
+          automountServiceAccountToken: false,
+          terminationGracePeriodSeconds: 1,
+          containers: [container],
+        },
+      } satisfies k8s.V1Pod;
+
+      await this.k8sClient.core
+        .createNamespacedPod(this.namespace, pod)
+        .catch(this.#handleK8sError.bind(this));
+    };
+
+    const createIdlePod = async (name: string, labels?: Record<string, string>) => {
+      const container = {
+        name,
+        image: "docker.io/library/busybox",
+        command: ["sh"],
+        args: ["-c", "sleep infinity"],
+      } satisfies k8s.V1Container;
+
+      await createPod(container, name, labels);
+    };
+
+    const createCompletedPod = async (name: string, labels?: Record<string, string>) => {
+      const container = {
+        name,
+        image: "docker.io/library/busybox",
+        command: ["sh"],
+        args: ["-c", "true"],
+      } satisfies k8s.V1Container;
+
+      await createPod(container, name, labels);
+    };
+
+    const createFailedPod = async (name: string, labels?: Record<string, string>) => {
+      const container = {
+        name,
+        image: "docker.io/library/busybox",
+        command: ["sh"],
+        args: ["-c", "false"],
+      } satisfies k8s.V1Container;
+
+      await createPod(container, name, labels);
+    };
+
+    await createIdlePod("test-idle-1", { app: "task-run" });
+    await createFailedPod("test-failed-1", { app: "task-run" });
+    await createCompletedPod("test-completed-1", { app: "task-run" });
+  }
+}