github.com/hatchet-dev/hatchet

refactor(cli): extract the profile store into a public package (#4942)

Merge branch 'main' into mk/try-assign-spans

ReplayTasks held one try-lock per tenant, so every replay for a tenant was
serialized through a single controller and any concurrent attempt failed
and cycled through the DLQ.

Lock each workflow run in the batch instead, with a blocking advisory lock
taken in ascending key order so overlapping batches cannot deadlock. The
batch still runs in one transaction. A concurrent replay of the same run
waits, then finds the tasks already queued and discards them in the
preflight check, so it degrades to a no-op rather than a duplicate replay.

Tasks without a workflow run id are locked on their own external id.

refactor(replay): replay each batch in one transaction, fetch inputs …

feat: disable the jit in more places

• fix: attempt to optimize concurrency queries

• fix: on conflict + retry handling

• fix: check retry count later

chore: add more spans inside try-assign-batch

• fix: cancelled due to concurrency should show as info log

• feat: pretty-print json-parseable error messages as json

• fix: open side panel for standalone tasks again

• fix: feature flag causing flashing

• chore: lint

• fix: handle error state

• fix: make all cancellation logs the same color

fix: attempt to optimize concurrency queries

• fix: optimize GetTenantTaskStats to try to make it less likely to seq scan v1_task

• fix: compiler

Merge remote-tracking branch 'origin/main' into belanger/serverless-o…

• fix: more gating of matches for operator spawns

• fix: rm unneeded skip

fix: oh boy.... (#4923)

Hatchet v0.106.5 is a concurrency-focused release, adding support for dynamic concurrency limits, shared concurrency across workflows, and two new queue-depth concurrency strategies.

Highlights

• Dynamic concurrency limits: set concurrency limits on a per-key basis using a CEL expression. For example, let premium-tier users run 10 workflows simultaneously while other users run one: input.tier == 'premium' ? 10 : 1 (#4863, #4873).
• Shared concurrency across workflows: create concurrency rules that many tasks and workflows consume simultaneously, e.g. limiting workflow_a and workflow_b to 10 concurrent runs per user total. Previously this pattern required a parent router task; shared concurrency replaces it (#4845, #4873).
• Two new concurrency strategies, CANCEL_QUEUED_EXCEPT_NEWEST and CANCEL_QUEUED_EXCEPT_OLDEST: configure the queue depth for each concurrency key, keeping only the most or least recent queued item until in-progress runs complete — conceptually similar to debouncing (#4793).

Security

• Engine: satisfied durable event lookups are now filtered by tenant, enforcing tenant isolation on the durable event log (GHSA-992g-9cr3-vm5x).

Fixed

• Engine: partitions are pruned in UpdateDurableEventLogEntriesSatisfied (#4891).
• Engine: CEL evaluation inconsistencies (#4864).
• Engine: duplicate rows no longer cause an on-conflict error during task assignment (#4829).
• Engine: allocated crons are counted against the latest workflow version only (#4859).
• Go SDK: all required symbols are exported and deprecated constructs are no longer depended on (#4875).

What's Changed

• feat: stamp cloud.region on OpenTelemetry resources by @grutt in #4780
• fix(engine): duped rows causing on conflict error on assign by @mrkaye97 in #4829
• refactor(engine): buffer durable event ingestion by @mrkaye97 in #4854
• chore(engine): decrease durable flush interval, add more spans by @mrkaye97 in #4858
• fix: count allocated crons via DISTINCT ON latest workflow version by @grutt in #4859
• docs: prioritize embedded mode by @abelanger5 in #4846
• refactor(engine): bulk operation for UpsertDurableChildSignalCreatedEvents by @mrkaye97 in #4860
• chore: Bump google.golang.org/grpc from 1.82.1 to 1.83.1 by @dependabot[bot] in #4866
• fix: dependabot alert by @mnafees in #4869
• fix: dependabot alert by @mnafees in #4870
• feat(engine): shared concurrency strategies by @abelanger5 in #4845
• feat: count ingested log and OTLP bytes after persist by @grutt in #4876
• feat(engine): enable durable tasks running on the dag operator to be evicted by @mrkaye97 in #4872
• chore: update skills to include embedded mode references by @abelanger5 in #4861
• feat(docs): add a reference overview page for each SDK reference by @abelanger5 in #4862
• fix: make sure Go SDK exports all required symbols and does not depend on deprecated constructs by @mnafees in #4875
• ci: no verbose go test output required by @mnafees in #4878
• chore: Bump qs from 6.15.2 to 6.16.0 in /frontend/app by @dependabot[bot] in #4877
• feat(engine): dynamic per-group max runs via CEL expression by @abelanger5 in #4863
• feat(dashboard): dag operator visibility by @mrkaye97 in #4881
• fix(python,ts): use a queue for ordering durable callbacks by @mrkaye97 in #4821
• feat(sdks): new concurrency features (dynamic max runs, tenant-scoped concurrency) by @abelanger5 in #4873
• fix: bumps qs in the TS SDK to fix Dependabot alerts by @mnafees in #4884
• chore: Bump github.com/rabbitmq/amqp091-go from 1.10.0 to 1.13.0 in /examples/go/embedded by @dependabot[bot] in #4885
• fix(python): concurrent workflow registration, flaky test improvement by @mrkaye97 in #4888
• ci: dont fail fast by @mrkaye97 in #4889
• fix: CEL inconsistencies / bugs in the engine by @abelanger5 in #4864
• feat(dashboard): add query params as span attributes on rest api requests by @mrkaye97 in #4892
• fix(engine): prune partitions in UpdateDurableEventLogEntriesSatisfied by @mrkaye97 in #4891
• fix: add new dag shapes to load test to stress test dag operator by @juliusgeo in #4857
• docs: link CEL feature clients on reference overview pages by @abelanger5 in #4893
• feat(dashboard): allow posthog flags to target users by email / id by @mrkaye97 in #4896
• feat(docs): index files w/ backlinks in the SDKs by @abelanger5 in #4895
• chore: pre-reqs for operators work (remove http operator, shared dispatcher types) by @abelanger5 in #4914
• chore: remove dependabot update PRs by @mnafees in #4919
• chore: Bump the github-actions group with 2 updates by @dependabot[bot] in #4906
• chore: Bump the npm-dev-deps group across 2 directories with 4 updates by @dependabot[bot] in #4835
• fix(engine): durable child spawn hang by @mrkaye97 in #4918
• fix: guard worker isActive using a session id instead of just a timestamp by @abelanger5 in #4916
• chore: Bump the go-modules group across 1 directory with 55 updates by @dependabot[bot] in #4921
• fix: Go SDK's RunMany should return results in the same order of the inputs by @mnafees in #4825
• chore: changelog for next release by @mnafees in #4915
• fix: correct OSS release tag by @mnafees in #4924

Full Changelog: v0.105.16...v0.106.5

• fix: hanging from child never spawned

• refactor: rm ReExecuted field on stuct

• chore: rm more dead code, rename struct

• fix: greptile concern

feat: allow posthog flags to target users by email / id (#4896)

• initial commit

• nested dag

• try this to fix issues

• fix started at incongruity

• fix race between dag starting and terminal statuses

• fix for stuck dags on engine restart

• ok try this

• ok try this

• alternative fix

• conditionally add to map

• fix other issues

• rewrite/remove bad comments

• add comment

• fix not null issue

fix(engine): prune partitions in `UpdateDurableEventLogEntriesSatisfi…

chore: Bump github.com/rabbitmq/amqp091-go in /examples/go/embedded (…

• no verbose go test output required

• remove atlas steps from CI

feat(engine): enable durable tasks running on the dag operator to be …

fix for stuck dags on engine restart

Merge branch 'main' into repro_load_test_dag

Merge branch 'main' into repro_load_test_dag

Merge branch 'main' into repro_load_test_dag

refactor(engine): bulk operation for `UpsertDurableChildSignalCreated…

• chore: add more spans to ingest

• chore: decrease flush interval

Merge branch 'main' into repro_load_test_dag