feat(webapp): mollifier trigger-time decisions — mollify, claim, read fallback by d-cs · Pull Request #3753 · triggerdotdev/trigger.dev

d-cs · 2026-05-26T10:43:48Z

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

`pnpm run typecheck --filter webapp` passes
`pnpm run test --filter webapp test/mollifierMollify.test.ts` passes
`pnpm run test --filter webapp test/mollifierIdempotencyClaim.test.ts` passes
`pnpm run test --filter webapp test/mollifierReadFallback.test.ts` passes
`pnpm run test --filter webapp test/mollifierGate.test.ts` passes
`pnpm run test --filter webapp test/engine/triggerTask.test.ts` passes

changeset-bot · 2026-05-26T10:43:53Z

⚠️ No Changeset found

Latest commit: 744a3da

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

coderabbitai · 2026-05-26T10:43:55Z

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 7843cb8d-9ccf-490a-a9a2-82c631270ed4

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch mollifier-phase-3-trigger

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

… fallback 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. - 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, parentTaskRun (triggerAndWait) skip the mollify path entirely. - triggerTask + IdempotencyKeyConcern wired to the above. Stacked on buffer extensions PR. All behaviour gated by the master `TRIGGER_MOLLIFIER_ENABLED` switch; off-state hot path is unchanged. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

`new Date("not-a-date")` returns a truthy Invalid Date object, which would mis-classify the run as CANCELED in the read-fallback synthesised shape. Add an `asDate` helper that rejects NaN-valued parses and use it for `cancelledAt` and `delayUntil`. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

The buffer's claim API now requires a caller-supplied ownership token so compare-and-act protects the slot against a stale predecessor. Wires the token end-to-end: - `claimOrAwait` generates the token (UUID) up front and reuses it across the retry path; returns it on the `claimed` outcome. - `publishClaim` and `releaseClaim` wrappers accept and forward the token to the buffer. - `ClaimedIdempotency` carries the token so the trigger pipeline can publish or release with the same token it claimed under. - `triggerTask.server.ts` threads the token into the publish call. Tests pin token round-trip: claimOrAwait → claimIdempotency, plus the publish and release pass-through. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…t leak fix) Devin review flagged a security + correctness bug: the mollify path returned \`MollifySyntheticResult\` with run shape \`{ friendlyId, spanId }\` and the call site cast it to \`TriggerTaskServiceResult\` (which expects \`run: TaskRun\` with an \`id: string\`). Downstream, the trigger route calls \`saveRequestIdempotency(requestKey, "trigger", result.run.id)\` — so the cache stored \`undefined\` as the entity id. On SDK retry the request-idempotency flow then ran \`prisma.taskRun.findFirst({ where: { id: undefined } })\`. Prisma strips \`undefined\` from where clauses, so the query degenerates to an unfiltered \`findFirst\` and returns an arbitrary TaskRun row — potentially from another env / user. Fix: - Add \`id: string\` to \`MollifySyntheticResult.run\`. - Compute it via \`RunId.fromFriendlyId(...)\` on both branches: the happy-accept path (id from \`args.runFriendlyId\`) and the \`duplicate_idempotency\` race-loser path (id from \`result.existingRunId\` so the response carries the WINNER's id). - Add regression tests pinning both branches. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…uard - readFallback: read snapshot.taskVersion (the key buildEngineTriggerInput writes) instead of the nonexistent snapshot.lockToVersion, so buffered version-locked runs report their locked version; test now uses the real key as a regression guard. - env: TRIGGER_MOLLIFIER_TRIP_THRESHOLD back to positive() (matching sibling mollifier numerics) to forbid threshold=0 silently mollifying every trigger. - idempotencyKeys: document why the resolved-but-unfindable fall-through is safe (PG-unique + accept SETNX dedup + ~30s claim TTL self-heal); add regression test pinning the fall-through and the resolved-and-findable cached-hit path. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Remove plan-tracking shorthand (Q#, C#, F#, Phase N, _plans/) from trigger-layer mollifier comments and test names; reword to plain English. Comment/test-name only; no behaviour change. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…f-state token alloc triggerTask.server.ts: mollifier.buffered logs at debug, not info — it fires per mollified trigger during a burst, exactly when info-level would flood aggregation. idempotencyClaim.server.ts: move the buffer-null fall-open check above token generation so the mollifier-OFF path skips the default randomUUID() for idempotency-keyed triggers (triggerTask is the highest-throughput path). Injected test generators still honoured. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

…nt, docs URL idempotencyKeys.server.ts: reword 'bypasses the gate via F4' to 'bypasses the gate entirely' — last residual internal planning-label reference I missed in the earlier cleanup pass. triggerTask.server.ts: reindent the startSpan arrow-function body to conventional depth (+4 sp on lines 143-619). The body was under-indented relative to the 'async (span) => {' opener after the try-wrapper for claim resolution was added around it; closes (arrow at 8sp, startSpan at 6sp) were already conventional, so only the body needed shifting. Pure whitespace — diff stat is balanced 413/413. mollifierMollify.server.ts: docs URL → https://trigger.dev/docs/management/tasks/batch-trigger. The previous burst-handling anchor doesn't exist on the docs site. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Addresses Devin's review on PR #3753 (idempotencyKeys.server.ts comment 2026-05-28): when TRIGGER_MOLLIFIER_ENABLED=1 globally during staged rollout, claimOrAwait was issuing a Redis SETNX for every idempotency-keyed trigger — including orgs that had not opted in. Those orgs gain nothing from the claim (the gate always returns pass_through for them, so the buffer is never written to) and PG unique constraint already deduplicates same-key races on the pass-through path. Wrap the claim block in an additional per-org check using the same in-memory Organization.featureFlags predicate that evaluateGate uses (makeResolveMollifierFlag) — no DB query. Non-opted-in orgs now skip claimOrAwait entirely; opted-in orgs still get the cross-store race coordination. Regression test added in mollifierClaimResolution.test.ts: with orgFlag=false the concern returns isCached:false with no claim, without calling claimIdempotency at all. The two prior tests now opt the fake org in (organization.featureFlags.mollifierEnabled=true) so they still exercise the resolution branches. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Addresses two Devin findings on PR #3753 (readFallback.server.ts:206 comments 2026-05-28). 1) batchId field-name mismatch: #buildEngineTriggerInput in triggerTask.server.ts writes batch info as a nested object 'batch: { id, index }', not as a flat 'batchId'. readFallback was reading the non-existent snapshot.batchId so SyntheticRun.batchId was always undefined for buffered runs. Now reads snapshot.batch.id (the internal cuid, matching what PG stores in TaskRun.batchId). 2) parentTaskRunFriendlyId / rootTaskRunFriendlyId structurally unfillable: the snapshot carries the INTERNAL parent/root ids (parentTaskRunId / rootTaskRunId, what engine.trigger consumes), not friendlyIds. Convert internal to friendly via RunId.toFriendlyId so consumers do not have to special-case the buffered path. Four regression tests in mollifierReadFallback.test.ts: batchId extracted from nested snapshot.batch.id; a flat snapshot.batchId key is ignored (belt-and-braces); parent/root friendly conversion round-trips through RunId.generate(); parent/root undefined when the snapshot has no parent context. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

d-cs commented May 26, 2026

View reviewed changes

Comment thread apps/webapp/app/v3/mollifier/readFallback.server.ts

d-cs self-assigned this May 26, 2026

d-cs force-pushed the mollifier-phase-3-trigger branch 2 times, most recently from 5a7bc19 to baa6f17 Compare May 26, 2026 13:24

This comment was marked as resolved.

Sign in to view

d-cs force-pushed the mollifier-phase-3-trigger branch from 01f3958 to 449a0bc Compare May 27, 2026 12:04

d-cs force-pushed the mollifier-phase-3-buffer branch from e85e771 to 9298706 Compare May 27, 2026 12:15

d-cs force-pushed the mollifier-phase-3-trigger branch from 449a0bc to ffe51b8 Compare May 27, 2026 12:15

d-cs force-pushed the mollifier-phase-3-buffer branch from 9298706 to f4c5b21 Compare May 27, 2026 12:21

d-cs force-pushed the mollifier-phase-3-trigger branch 6 times, most recently from cae33fa to 16bfff0 Compare May 27, 2026 16:50

d-cs marked this pull request as ready for review May 28, 2026 09:03

0ski approved these changes May 28, 2026

View reviewed changes

d-cs and others added 9 commits May 28, 2026 10:40

docs(server-changes): document mollifier trigger-time decisions

bc274cf

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

d-cs force-pushed the mollifier-phase-3-trigger branch from f126737 to 36cc024 Compare May 28, 2026 09:41

This comment was marked as resolved.

Sign in to view

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat(webapp): mollifier trigger-time decisions — mollify, claim, read fallback#3753

feat(webapp): mollifier trigger-time decisions — mollify, claim, read fallback#3753
d-cs wants to merge 11 commits into
mollifier-phase-3-bufferfrom
mollifier-phase-3-trigger

d-cs commented May 26, 2026 •

edited

Loading

Uh oh!

changeset-bot Bot commented May 26, 2026 •

edited

Loading

Uh oh!

coderabbitai Bot commented May 26, 2026 •

edited

Loading

Review skipped

Uh oh!

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

d-cs commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test plan

Uh oh!

changeset-bot Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

⚠️ No Changeset found

Uh oh!

coderabbitai Bot commented May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review skipped

Uh oh!

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

d-cs commented May 26, 2026 •

edited

Loading

changeset-bot Bot commented May 26, 2026 •

edited

Loading

coderabbitai Bot commented May 26, 2026 •

edited

Loading