Commit Graph
2020 Commits
Author SHA1 Message Date
5f8e0022b6 feat: fairness reserve as a worker group setting, shown on the workers page (#12435)
* feat: fairness reserve as a worker group setting, shown on the workers page

Replaces the WORKSPACE_FAIRNESS_RESERVE_WORKER env var with a
`fairness_reserve` worker group setting, toggled in the group config
editor and shown as a badge on the workers page. Workers pick it up on
config reload without restarting. A worker still started with the env
var exits with a message pointing to the setting.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: only refuse the old reserve env var when it enabled the flag

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvJe56H5vnK7jXVvAp

* fix: warn that a fairness reserve idles when fairness is off, at startup and on enable

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvJe56H5vnK7jXVvAp

* fix: warn on enabling a fairness reserve while fairness is off

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvJe56H5vnK7jXVvAp

* feat: log feature usage for the fairness reserve toggle

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvJe56H5vnK7jXVvAp

* chore: update ee-repo-ref to 3972beb0b39a115aad0154d43558692627158b3b

This commit updates the EE repository reference after PR #986 was merged in windmill-ee-private.

Previous ee-repo-ref: 15a06f4c8dff6e8d8fb4cc92ce1655bb48a41b02

New ee-repo-ref: 3972beb0b39a115aad0154d43558692627158b3b

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-14 16:17:31 +01:00
Ruben FiszelandClaude Opus 5.5 df56c111bb feat: fairness reserve workers that only serve non-overloaded workspaces (#12431)
* feat: fairness reserve workers that only serve non-overloaded workspaces

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse the fairness reserve flag on agent workers

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: let reserve workers resume suspended jobs of non-overloaded workspaces

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refresh overload state before reserve resumes, share suspended query

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse reserve workers on CE builds, warn while fairness is off

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: gate the reserve flag on the enterprise feature, tidy comments

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: fairness setting comment names the enterprise gate

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-14 07:55:44 +01:00
3b860d549e fix: keep step_id, job_id, flow_path and extra on dedicated worker errors (#12417)
* fix: keep step_id, job_id, flow_path and extra on dedicated worker errors

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: add extra to python and step context to native dedicated worker errors

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep dedicated python worker alive on unserializable error attributes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 07883025e535974b9a01b20913f904b8f798d311

This commit updates the EE repository reference after PR #982 was merged in windmill-ee-private.

Previous ee-repo-ref: 129aa6a625084ea3a35d8f37bf3b825f1b2cee26

New ee-repo-ref: 07883025e535974b9a01b20913f904b8f798d311

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-13 17:36:58 +01:00
hugocasaandwindmill-internal-app[bot] f738a4bad4 fix: run dedicated flow script steps on their runners (#12419)
* fix: run dedicated flow script steps on their runners

* fix: give pinned script versions their own runner directory

* chore: update ee-repo-ref to 3162714b4eeafb33058eab075f9c24403fc7f690

This commit updates the EE repository reference after PR #983 was merged in windmill-ee-private.

Previous ee-repo-ref: 4d2b5c5ec67f86e54178f314809e8ac9a404c061

New ee-repo-ref: 3162714b4eeafb33058eab075f9c24403fc7f690

Automated by sync-ee-ref workflow.

---------

Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-13 17:34:56 +01:00
c93a17396c feat: vertex_ai_federated resource type for workload identity federation (#12416)
* feat: Workload Identity Federation for Vertex AI resources

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: list vertex_workload_identity in feature telemetry

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: vertex_ai_federated resource type instead of fields on anthropic and googleai

A federating resource carries no key and only works where Windmill runs the
exchange, so it gets its own type: the anthropic and googleai types keep a
required key and their generated script types stay unchanged. The type implies
Vertex, and the pickers infer Claude or Gemini from its base_url publisher.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: resolve a variable base URL, keep federated Vertex out of decisions, drop stale picks

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: point token-based Vertex AI resources at vertex_ai_federated

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 458eeca8c31ba093a72eeb1ed9d82a83e18477db

This commit updates the EE repository reference after PR #981 was merged in windmill-ee-private.

Previous ee-repo-ref: 269785e57e681bea9be0c66ca1b8ba92b04d836a

New ee-repo-ref: 458eeca8c31ba093a72eeb1ed9d82a83e18477db

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-13 15:40:06 +01:00
Ruben FiszelandClaude Opus 5.5 b2793f2aca fix: stop logging credential-bearing instance setting values (#12413)
* fix: stop logging string and object instance setting values

The generic setting loaders printed every value loaded from global_settings,
so credentials in bunfig_install_scopes, npmrc, pip_index_url,
powershell_repo_pat, cargo_registries, maven_settings_xml, hub_api_secret, ...
landed in clear in server and worker logs at startup and on every reload.
Only scalar values are printed now; strings, arrays and objects are reduced
to their length.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep registry tokens and index urls out of worker debug logs

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-13 14:34:42 +01:00
792e66a30a feat: one turn per memory id, a new message takes over a parked turn (#11831)
* feat(ai-agent): background tools that wake the agent when they finish

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): run a tool in the background after a delay

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): one running turn per memory id, take over, keep history on stop

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): claim the memory id of an app's agent run

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): claim an app run's memory id after its conversation check

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore(ai-agent): date the memory turn migration after main's

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): a new run takes over a parked turn of its memory id

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(flow-chat): send while the turn is parked

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs(ai-agent): a parked turn hands its memory id over

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor(flow-chat): let the server decide whether a turn is parked

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor(ai-agent): memory turn queries as sqlx macros

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor(ai-agent): park and background tool queries as sqlx macros

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(flow-chat): show background tool calls as one running card, no typing dots while waiting

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): hand a turn over only from a run's last agent step

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(chat-sdk): Stop cancels a message still sent over a running turn

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(flow-chat): say when a call moved to the background and when results came back

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(chat-sdk): one message over a running turn at a time

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(flow-chat): one card per background call, without the model's note as its result

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(flow-chat): a background result streamed in early still folds its row into the call's card

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(flow-chat): queue a message over a working turn without asking the server, send it once the turn waits

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(ai-agent): a handed-over background job follows its new turn's cancel; retry a send refused just before a park

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(flow-chat): a stopped background call stays finished in later turns, and a late row read keeps a streamed result

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): a hand-over keeps the old turn's background calls open even after their rows were read

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): a late background row keeps a streamed failure failed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(ai-agent): say the agent waits once its memory is saved; cancel handed jobs no turn adopted instead of re-parenting them early

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(ai-agent): cancel unadopted handed jobs with their descendants; a handed-over turn stopped mid-resume saves no memory

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): no other send while one sent over a turn waits, even once that turn ended

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* chore: restore sqlx cache entries the main merge dropped

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): a background result row stays with its own call when the tool is called again; drop stale sqlx entries

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-13 13:21:40 +01:00
hugocasaandClaude Opus 5.5 10a8577fab feat: background tools for ai agent steps (#11827)
* feat(ai-agent): background tools that wake the agent when they finish

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): run a tool in the background after a delay

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor(ai-agent): park and background tool queries as sqlx macros

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(flow-chat): show background tool calls as one running card, no typing dots while waiting

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(flow-chat): say when a call moved to the background and when results came back

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(flow-chat): one card per background call, without the model's note as its result

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(flow-chat): a background result streamed in early still folds its row into the call's card

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(flow-chat): a stopped background call stays finished in later turns, and a late row read keeps a streamed result

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): a late background row keeps a streamed failure failed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* chore: restore sqlx cache entries the main merge dropped

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

* fix(chat-sdk): a background result row stays with its own call when the tool is called again; drop stale sqlx entries

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F4rzFpemqD4pqLnAp4qRqF

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-13 13:12:29 +01:00
70711bec68 feat: mount volumes read-only without the exclusive lease (#12410)
* feat: mount volumes read-only without the exclusive lease

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvmmTkuELF5qpS2Hs3

* fix: gate lease-free volume reads on the fleet version

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvmmTkuELF5qpS2Hs3

* fix: keep held volumes counted by cache eviction, restore MIN_VERSION in tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvvmmTkuELF5qpS2Hs3

* chore: update ee-repo-ref to 38904055ab0d104006278322299bdbcea32c88d1

This commit updates the EE repository reference after PR #980 was merged in windmill-ee-private.

Previous ee-repo-ref: 1fe25fa2530efdba4f8a3bf5dbfbb6f85e616ce7

New ee-repo-ref: 38904055ab0d104006278322299bdbcea32c88d1

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-13 07:55:54 +01:00
d4f9975121 fix: skip the worker's suspend_resume audit row when a rejected approval continues (#12376)
The API already records the rejection under the approver; the worker added a
second row under the flow's requester saying the step timed out.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-12 12:40:11 +01:00
Ruben FiszelandClaude Opus 5.5 e9f4636d44 feat: run AI decision steps on Gemini's Decisions API (#12370)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-12 07:32:28 +01:00
Ruben FiszelandClaude Opus 5.5 bd5f4d4d3d fix: run ci tests once a batch deploy has landed, not per item (#12369)
* fix: run ci tests once a batch deploy has landed, not per item

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: require deploy rights for ci test batches and wait out delayed relocks

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: gate ci test batches on run role and keep failed pushes tested

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-12 07:27:38 +01:00
AlexRV12andwindmill-internal-app[bot] d78bc9b026 refactor: say who chose the path wherever a job is built from one (#12358)
* refactor: say who chose the path wherever a job is built from one

* refactor: rename SavedItem to Deployed, checked when saved

* refactor: trigger pushes are deployed, test connection named by the tester

* chore: update ee-repo-ref to bf04cefe64776cb7262f041105b6ceca2953d5f4

This commit updates the EE repository reference after PR #976 was merged in windmill-ee-private.

Previous ee-repo-ref: 2276bc7f7f711d6ef33a5133ca87e78ae0fa345a

New ee-repo-ref: bf04cefe64776cb7262f041105b6ceca2953d5f4

Automated by sync-ee-ref workflow.

---------

Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-12 07:16:25 +01:00
hugocasaandClaude Opus 5.5 9d9875cf0e fix: treat skipped jobs as successful in result and wait paths (#12340)
* fix: treat skipped jobs as successful in result_maybe, run-and-wait, AI tools and Slack agents

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: return is_skipped from get_result_maybe

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-11 12:24:58 +01:00
hugocasaandClaude Opus 5.5 5b4f918e0d fix: stop forcing the structured output tool for Claude AI agents (#12331)
* fix: use native structured outputs for Claude AI agents, not a forced tool

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep the structured output tool for Claude but stop forcing it

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 23:02:01 +01:00
Ruben FiszelandClaude Opus 5.5 e3ff8bee3d fix: never break running flows when their flow is deleted (#12306)
* fix: report the real error when a flow's stored version cannot be loaded

A flow step that completes on a worker without the flow version cached
looked it up, failed (e.g. the flow was deleted, cascading its versions),
then fell back to the preview path, whose 'Job isn't a flow job' error
replaced the real one. Keep the stored lookup's error when the job has no
inline flow.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a flow's nodes when it is renamed

Renaming a flow moved its versions to the new path but left its flow nodes
on the old one, so deleting the old flow row cascaded them away. Runs in
flight inside a loop or branch then failed loading their body.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a deleted flow's versions until its runs are done

Deleting a flow cascaded its flow_version and flow_node rows away, so runs
in flight failed at their next step. The cascade is dropped; the rows are
purged when the flow leaves the trash, unless a run of it is still queued,
in which case they are kept. Workspace deletion deletes them explicitly,
and the version history only lists the current flow's versions.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: scope moves, reads and the purge to a flow's own versions

A deleted flow's versions now survive at its path, so rename, user rename,
offboarding, fork cloning, version reads, history and the trash purge only
act on the ids the live flow lists. Renames skip nodes already present at the
destination with the same content, nodes are no longer purged (they are
shared by content and referenced by id), and dependency relocks skip flows
that no longer exist.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: resolve a flow's latest version among the live flow's own

The relock checks and the dependents trigger picked the newest version at the
path, which can now be a deleted flow's. Restore the shared query's cache entry
under its new text.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: set a deleted flow's versions aside instead of scoping every lookup

Deleting a flow moves its versions to __deleted/<trash id>; restoring moves
them back. Jobs load versions by id so runs in flight keep working, while
every lookup by path (history, relocks, versioned runs, renames) stops
seeing a deleted flow without special-casing. The purge keeps a set-aside
version any queued job references by id, whatever path or kind it was
queued under. The per-lookup scoping is reverted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* style: rustfmt

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: set versions aside only after an authorized delete, find flow_env by id

Row-level security can let a caller read a flow it may not delete; the
delete then removes nothing and must not trash it or move its versions.
fetch_root_flow_env matched a run's version by path too, so runs of a
deleted (or renamed) flow lost their inherited flow_env: match by id for
the kinds whose runnable_id is a flow version.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: state the authorization contract of the version set-aside helpers

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* perf: read the queue once when purging a deleted flow's versions

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: serialize flow version writes with a concurrent flow delete

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: make the relock copy wait for a concurrent flow rename

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 10:23:10 +01:00
Ruben FiszelandClaude Opus 5.5 3bf486415b fix: keep flow nodes on rename and stop masking flow load errors (#12303)
* fix: report the real error when a flow's stored version cannot be loaded

A flow step that completes on a worker without the flow version cached
looked it up, failed (e.g. the flow was deleted, cascading its versions),
then fell back to the preview path, whose 'Job isn't a flow job' error
replaced the real one. Keep the stored lookup's error when the job has no
inline flow.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a flow's nodes when it is renamed

Renaming a flow moved its versions to the new path but left its flow nodes
on the old one, so deleting the old flow row cascaded them away. Runs in
flight inside a loop or branch then failed loading their body.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-09 22:51:39 +01:00
04230c286d feat: reload dedicated workers in place and restart them one at a time (#12208)
* feat: reload dedicated workers in place and restart them one at a time

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: reload dedicated importers on deploy, refresh python imports, count native readiness

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: count a native dedicated worker as started once its isolate is ready

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: explain staggered restarts in the dedicated worker group config

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: state when dedicated deploys restart workers in the group config copy

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: shorten the dedicated worker group restart copy

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: warn about staggered restarts when applying a dedicated group config

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show the staggered restart note only for groups with several workers

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: say when a dedicated deploy restarts workers in the group config note

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: clarify when dedicated deploys restart workers and tighten the dedicated tab spacing

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep only the supervisor note in the dedicated tab

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: word the staggered restart alert in terms of workers

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: use checked sqlx macros for the dedicated restart turn queries

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: list what restarts dedicated workers in a tooltip on the group config

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: simplify the dedicated deploy restart bullet

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 51a806ae14dcca95ca2221a9f5cc6434ed755bca

This commit updates the EE repository reference after PR #950 was merged in windmill-ee-private.

Previous ee-repo-ref: ce1eba465e5891bab71173c06723a1a6cbd2d869

New ee-repo-ref: 51a806ae14dcca95ca2221a9f5cc6434ed755bca

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-09 17:26:54 +01:00
hugocasaandClaude Opus 5.5 e6183f5e0f fix: mint an RDS IAM token for DuckDB postgres attaches (#12260)
* fix: mint an RDS IAM token for DuckDB postgres attaches

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: drop unrelated module reorder

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mint the token for a fork's ancestor ducklake catalogs too

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: state the caller contract of iam_auth_token

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mask longer secrets first and skip empty ones in duckdb errors

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mask secrets in the recorded materialization error

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-09 17:10:45 +01:00
AlexRV12andwindmill-internal-app[bot] a70afdea36 refactor: shared resolution helpers take a RunnableRef (#12285)
* refactor: shared resolution helpers take a RunnableRef

* chore: update ee-repo-ref to 6e0c858148b7099399be3f9a0ab81b258ca88dac

This commit updates the EE repository reference after PR #968 was merged in windmill-ee-private.

Previous ee-repo-ref: be7a939d7ba0ad78ebc5a47ab130f5677c5f5d8e

New ee-repo-ref: 6e0c858148b7099399be3f9a0ab81b258ca88dac

Automated by sync-ee-ref workflow.

---------

Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-09 05:18:50 +01:00
Alexander PetricandClaude Opus 5.5 e8034a1cf5 fix: warn instead of error on isolation probe failure when not requested (#12288)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-09 05:16:26 +01:00
Ruben FiszelandClaude Opus 5.5 728b3dd1aa fix: install ansible python deps for the python ansible-playbook runs on (#12283)
* fix: install ansible python deps for the python ansible-playbook runs on

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: log when an ansible python lock does not match the ansible python

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvwanCe5xnmjd2Bcv2b

* fix: skip the ansible python version warning for a lock without pins

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvwanCe5xnmjd2Bcv2b

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-06 17:22:27 +00:00
Ruben FiszelandClaude Opus 5.5 caf0d67563 feat: track failed job rates per worker (24h, last 100, lifetime) (#12275)
* feat: track failed job rates per worker over 24h, last 100 jobs and lifetime

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: count pre-execution failures and skip pre-canceled jobs in worker stats

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: persist worker job stats after the result processor drains

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: report an agent's failed init script in its job stats, document what is counted

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show one failure rate per worker row, the other windows in its tooltip

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show worker failure rate as a bar under jobs ran instead of a column

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show group failure rate as a bar on the active workers line

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: align the group failure label on the active workers baseline

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show the group's jobs ran count next to its failure bar

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-06 17:04:27 +01:00
Ruben FiszelandClaude Opus 5.5 2516d3fca2 fix: apply preserve_step_tags to steps of untagged subflows (#12278)
* fix: apply a flow's preserve_step_tags to steps of its untagged subflows

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: settle enclosing preserve_step_tags once per push and for agent tools

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: stop at subflows that declare their own tag when inheriting preserve_step_tags

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: walk preserve_step_tags from the flow's own tag and skip default flow tags

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-06 16:40:04 +01:00
Ruben FiszelandClaude Opus 5.5 8a53808622 fix: never relock a codebase script into a version without its bundle (#12271)
* fix: never relock a codebase script into a version without its bundle

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: say which layer keeps codebase scripts from being relocked

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a plain redeploy's importer rows when fan-out cleanup races it

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: leave a codebase importer's rows alone in the relock fan-out

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-06 13:26:47 +01:00
AlexRV12 2b95c9a73e fix: parse flow step, ai tool and wac step paths on the worker (#12264) 2026-11-06 11:13:57 +00:00
964f1af303 feat: option to fail a run on a full concurrency key (#12215)
* feat: fail or queue on a full concurrency key, keep queue order on requeue

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mint a flow token for pre-run step failures, accept on_concurrency_limit in ai flow settings

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: fail reserved ai tools on a full concurrency key in fail mode

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: drop ordered re-queue, keep the fail mode

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 0b763cc905bb67b17728794bb52bfdf336fe25fb

This commit updates the EE repository reference after PR #952 was merged in windmill-ee-private.

Previous ee-repo-ref: 9e92c30589c417c5eae156b28760470ad4e86628

New ee-repo-ref: 0b763cc905bb67b17728794bb52bfdf336fe25fb

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-05 23:31:06 +01:00
hugocasaandClaude Opus 5.5 9cd04d0a88 fix: keep the real error when a Bun script throws a cyclic error (#12232)
* fix: keep the real error when a Bun script throws a cyclic error

The Bun wrapper copies a thrown error's own properties into `extra` and
JSON-encodes the result. When one of them is cyclic (AWS SDK v3 errors'
`$response`), the encoding threw and the job error became
"cannot serialize cyclic structures". Fall back to an encoding that
replaces back-references with "[Circular]", keeping the plain encoding
first so non-cyclic errors are unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: replace only the unencodable fields of a Bun error's extra

Encode each field of `extra` on its own when the whole error fails, and
replace the ones that throw with "[unserializable]". Simpler than walking
cycles, and drops noise like AWS's `$response` while keeping `Code` and
`$metadata`.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-05 23:29:19 +01:00
c057f8995d feat: optional reason when denying an approval (#12213)
* feat: optional reason when denying an approval

* fix: keep the full deny reason in the audit entry

* fix: name the denial reason in the payload guard and document error.reason

* feat: tuck the deny reason behind an add a reason link

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: ask for the deny reason in a popover on deny

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: label the run page cancel action cancel step

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-05 23:27:11 +01:00
Ruben FiszelandClaude Opus 5.5 3c5a5caf97 feat: inherited low priority for background flows and scripts (#12256)
* feat: add inherited low priority for background flows and scripts

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a low flow priority over step priorities and align the -1 boundary

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: note the rolling-upgrade ordering and the renamed queue index

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: order low priority by stored values instead of a coalesce index

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a runnable's own priority when its caller has none

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: do not pass a flow step's default priority on to the jobs it starts

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-05 23:26:04 +01:00
b52eb8b6c2 feat: cap concurrent branches of a parallel branch-all step (#12211)
* feat: cap concurrent branches of a parallel branch-all step

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: reject a zero parallelism cap and regenerate flow guidance

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-05 23:03:46 +01:00
0ab7290f67 feat: let flow step mocks apply to preview runs only (#12243)
* feat: mark mocked and skipped steps in the flow run view and step logs

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep the skipped badge to what the run state can tell

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: skip the mocked badge on expanded subflow steps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: keep only the step log lines for mocked and skipped steps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: let flow step mocks apply to preview runs only

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: apply the test-runs pin default when pinning from the flow graph

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: choose a pin's scope from the pin and show pins where they apply

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: build the pin scope menu on the standard split button

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-05 22:58:52 +01:00
hugocasaandClaude Opus 5.5 4c9959ba1a feat: log in a mocked or skipped flow step that nothing was executed (#12212)
* feat: mark mocked and skipped steps in the flow run view and step logs

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep the skipped badge to what the run state can tell

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: skip the mocked badge on expanded subflow steps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: keep only the step log lines for mocked and skipped steps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-05 20:05:05 +01:00
71beff7438 feat: add max iterations and max duration bounds to while loops (#12245)
* feat: add max iterations and max duration bounds to while loops

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: fail exhausted while loops regardless of skip_failures and prior recovery

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: state that a while loop's duration limit can overrun by one iteration

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: let the while loop iterations field be cleared and retyped

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-05 15:15:03 +00:00
Ruben FiszelandClaude Opus 5.5 1f134e4127 fix: honor iam and azure workload identity on server-side pg connections (#12228)
* fix: honor iam and azure workload identity on server-side pg connections

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse token auth modes for resource sources on the server

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: send the server's entra token only to verified azure pg hosts

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse a token-mode clone before creating its database

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-05 09:41:27 +01:00
f9448bf25b feat: filter the audit log by run (#12190)
* feat: filter the audit log by run

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: say a nested approval lists under its subflow

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: put the run's audit logs in a menu on the left of its actions

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* perf: open a run's audit logs on the run's own time range

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: include the run's steps and subflows in the run filter

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: name each step and mark mocked or skipped steps in run audit entries

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show a step's id in the resource cell instead of next to the operation

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep step details in the audit drawer and show its parameters as json

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 0bb5b777e634dadd154c0adfe6d26251560f9283

This commit updates the EE repository reference after PR #946 was merged in windmill-ee-private.

Previous ee-repo-ref: b2f1890730a3368019183985161744187b5323a7

New ee-repo-ref: 0bb5b777e634dadd154c0adfe6d26251560f9283

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-11-04 20:34:23 +01:00
6268780952 fix: stop git sync auto-pull from reverting edits with its own commit (#12209)
* fix: stop auto-pull from reverting edits with windmill's own git sync commit

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: say when should_pull skips a commit windmill pushed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: skip the synced sha advance when a pull overlaps the push

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test: cover an overlapping pull from another hub version

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: pin the git sync push script to hub 30168

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 30ca75f61bd354d18c9edc7fb0045f7d69480d64

This commit updates the EE repository reference after PR #951 was merged in windmill-ee-private.

Previous ee-repo-ref: f9427c628e2e81c6d3e16ede7e384a5bafc69a8a

New ee-repo-ref: 30ca75f61bd354d18c9edc7fb0045f7d69480d64

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-11-04 20:18:57 +01:00
hugocasaandClaude Opus 5.5 6a1a0a3599 feat: flow-level wall-clock timeout (#12205)
* feat: flow-level wall-clock timeout

* chore: regenerate flow schema prompts after rebase

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: claim timed-out flows before cancelling, scope the handler to the expired flow

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: scope the flow timeout to root flows

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: drop flownode from the flow timeout sweep

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: apply a flow's timeout when it runs as a subflow step

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: align subflow timeouts with script timeouts and let the ai chat set them

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: give a calling step the timeout error whether or not the subflow has a failure module

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 18:51:43 +01:00
Ruben Fiszel 116c90280e feat: name a materialized table after its script when the target names none (#12157)
* fix: teach pipeline nodes to attach the lake and read real columns

* test: run the data pipelines ai evals build against a real ducklake

* test: run only fed-by-nothing pipeline nodes and record run failures

* feat: name a materialized table after its script when the target names none

* fix: inspect pipeline columns read-only and cap the eval run's roots

* test: cap the roots of the tables-named-after-steps eval

* fix: keep flow steps and explicit targets out of the materialize default

* fix: name the table of a script run from a flow, recover bare materialize in the cli

* ci: point eval workers at the installed bun and serve native scripts

* fix: keep inline step tests out of the materialize default, recover bare scd2 in the cli

* chore: drop formatter-only changes from the materialize default diff

* fix: let a referenced-script step test name its table after the script

* fix: tell inline step tests apart by the missing script hash, not the path
2026-11-03 03:38:51 +01:00
Ruben FiszelandClaude Opus 5.5 3d0ca92ef2 feat: upgrade the bundled DuckDB engine to 2.0.0 LTS (#11924)
* feat: upgrade the bundled DuckDB engine to 2.0.0 LTS

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: never repeat a ducklake attach option, duckdb 2.0 rejects duplicates

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: dedupe quoted attach option names and drop comments from the list

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: pin duckdb 2.0 from crates.io and revert more 2.0 defaults that change results

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: fail duckdb jobs with a clear message on x86 cpus without avx2

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: strip sql comments from lake attach args before appending fork options

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: scan E'' escape-string literals when stripping attach option comments

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-02 15:40:22 +01:00
45f256a5a1 feat: accept MCP access tokens from an external OIDC issuer (#12085)
* feat: accept MCP access tokens from an external OIDC issuer

With the mcp_external_idp instance setting, every MCP URL also accepts an
access token of an external OIDC issuer, told apart from Windmill's own
by its iss. The token's OAuth client is mapped to a workspace and a
selection of tools; the request acts as the existing Windmill user with
the token's email, under that selection. Runs record the client in
mcp_client and give it to steps as WM_MCP_CLIENT; tool calls and
refusals are audited with both identities.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: harden external MCP token verification

Refuse ID tokens, tokens before their nbf, and emails the issuer marks
unverified; let a cached signing key expire so a withdrawn key stops
verifying; fall back to the token's asymmetric alg for keys published
without one; reject clients configured without an issuer or with an id
longer than the column recording it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse deactivated accounts and untyped tokens on external MCP auth

Read password.disabled, since instance deactivation has no Windmill
token to revoke on this path. Without a configured audience, take only
tokens typed as access tokens (at+jwt header or a Bearer typ claim), so
an ID token cannot stand in for one. Require https for the issuer except
on localhost.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: refuse ID tokens with an audience set, parse the issuer host

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: take an external MCP token's groups for the request instead of storing them

When sso_groups_claim is set, the IdP-managed instance groups of an
external MCP request come from the token's claim alone, mapped through
their external_id with nested paths counting for their ancestors. Nothing
is written, and an absent claim means no IdP groups, so removing a user
from every group revokes at the next token. Entra's overage pointer is
refused rather than read as no groups.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: give external MCP clients the token access levels, one card per client

A client takes the same access as a Windmill token: read, read and run,
build and run (the session toolset, bound to the client's workspace), or
scripts and flows as tools. The settings editor reuses McpAccessPicker,
one collapsible card per client. The caller's JWT carries mcp_client so
runs the session toolset starts record the client too. With token groups
in use, admin rights a managed group grants follow the token, and
workspace groups are kept by source rather than by name.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep a client's tool selection across reopening its card

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: tidy the MCP external identity provider form

Show it under SSO only, shorten its description with details in a
tooltip, and give its fields form labels, hints and tooltips.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: leave default claim names empty, shorten auth setting descriptions

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: report an external MCP token's groups in whoami

Requests and jobs started through an external MCP token act with the
token's groups; whoami returned the stored memberships instead.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep stored roles for external MCP requests

Only group memberships follow the token. Admin and superadmin roles stay
as stored, kept current by SCIM and SSO login: their source is not
recorded, so a group-derived demotion also stripped directly granted
admins and left operator restrictions unapplied.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: record the MCP client on on-behalf-of runs and AI agent tool steps

On-behalf-of runs now hand push the MCP caller, which push uses only to
record the client while it fetches the run's own permissions. AI agent
tool jobs read the client from job_perms, so their steps see
WM_MCP_CLIENT.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep the MCP client off v2_job

The client is needed only while a job is queued, which job_perms covers; a
finished run's client is in its audit entry. This leaves the job history
table unaltered.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to e9e94030cebd5afb0c8156ee46af2025bbe8778a

This commit updates the EE repository reference after PR #934 was merged in windmill-ee-private.

Previous ee-repo-ref: 7bdd6e8067a72e847ed0372ddc79a84f288946cf

New ee-repo-ref: e9e94030cebd5afb0c8156ee46af2025bbe8778a

Automated by sync-ee-ref workflow.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
2026-10-30 20:09:09 +01:00
hugocasaandClaude Opus 5.5 37b1ea23f9 fix: keep secret values out of stored step inputs from flow_env (#12130)
* fix: keep secret values out of stored step inputs reached through flow_env

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: pass secret flow_env entries to subflow steps as references

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test: cover a subflow input computed from a secret flow_env entry

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: detect a secret flow_env read only through a flow_env access

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: pass a subflow input computed from a secret flow_env entry encrypted

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* perf: load flow_env secrets only for an input that reads flow_env

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 18:17:35 +01:00
AlexRV12 6df5705223 fix: run Hub script trigger targets that have a retry or error handler (#12094) 2026-10-29 19:25:41 +01:00
Ruben FiszelandClaude Opus 5.5 d03142ba39 perf: borrow cached script modules instead of cloning per job (#12087)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 21:45:27 +01:00
hugocasaandClaude Opus 5.5 e480501af9 feat: accept the google_service_account resource type alongside gcp_service_account (#12070)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 19:46:04 +01:00
hugocasaandClaude Opus 5.5 499c7569ff fix: honor a step's timeout when it is the only step of a loop (#12058)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 13:37:14 +01:00
375bf03cdc feat: review git sync pull requests with ai on the instance (#12056)
* feat: add ai review to deployment requests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: bound ai review fetches and surface unreadable items

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: add fix in ai session and model picker to deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: add workspace instructions for deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: read review instructions from the parent and bind ai review to the fork

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: pin the fix-all session preview to the fork's compare page

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show the protection-rule bypass toggle inline

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: make the bypass toggle small and red

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: drop workspace-settings instructions from deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: folder-scoped review instructions for the deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: review the selected items with ai on compare and deploy

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep reviewing when review instructions cannot be listed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: say the review instruction floor is a prompt request, not enforcement

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mark an ai review out of date when its items change

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 77703cf32e58964906d87ee83ddda44b5b8d1b7e

This commit updates the EE repository reference after PR #926 was merged in windmill-ee-private.

Previous ee-repo-ref: e30ab9e69d1854fcaed5620c934ace32acb0a729

New ee-repo-ref: 77703cf32e58964906d87ee83ddda44b5b8d1b7e

Automated by sync-ee-ref workflow.

* feat: review git sync pull requests with wmill review

* fix: withhold retyped resources and encryption keys, fall back from the responses route

* fix: withhold fileset children, workspace-specific resources and pathspec-named files

* feat: include the ready-to-post review in wmill review --json

* feat: review a workspace synced into a repository subfolder

* docs: say what wmill review takes from flags rather than wmill.yaml

* feat: review git sync pull requests with ai on the instance

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvwbvQh6Traiw1J3fRM

* chore: update ee-repo-ref for the ai review fixes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvwbvQh6Traiw1J3fRM

* feat: point git sync ai review at the published hub script

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PjsYvwbvQh6Traiw1J3fRM

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-10-28 06:05:47 +01:00
fbe776d611 feat: review git sync pull requests with wmill review (#12055)
* feat: add ai review to deployment requests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: bound ai review fetches and surface unreadable items

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: add fix in ai session and model picker to deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: add workspace instructions for deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: read review instructions from the parent and bind ai review to the fork

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: pin the fix-all session preview to the fork's compare page

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: show the protection-rule bypass toggle inline

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: make the bypass toggle small and red

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* refactor: drop workspace-settings instructions from deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: folder-scoped review instructions for the deployment ai review

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat: review the selected items with ai on compare and deploy

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep reviewing when review instructions cannot be listed

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* docs: say the review instruction floor is a prompt request, not enforcement

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: mark an ai review out of date when its items change

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore: update ee-repo-ref to 77703cf32e58964906d87ee83ddda44b5b8d1b7e

This commit updates the EE repository reference after PR #926 was merged in windmill-ee-private.

Previous ee-repo-ref: e30ab9e69d1854fcaed5620c934ace32acb0a729

New ee-repo-ref: 77703cf32e58964906d87ee83ddda44b5b8d1b7e

Automated by sync-ee-ref workflow.

* feat: review git sync pull requests with wmill review

* fix: withhold retyped resources and encryption keys, fall back from the responses route

* fix: withhold fileset children, workspace-specific resources and pathspec-named files

* feat: include the ready-to-post review in wmill review --json

* feat: review a workspace synced into a repository subfolder

* docs: say what wmill review takes from flags rather than wmill.yaml

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
2026-10-28 05:49:07 +01:00
hugocasaandClaude Opus 5.5 a1bc9ec0c2 feat: tool approval and a question tool for ai agent steps (#11826)
* feat(ai-agent): built-in approval tool that parks the agent

An AI agent step can now be given a built-in "Request approval" tool. The model
calls `request_approval` with a message; the agent stores its conversation in
the new `ai_agent_checkpoint` table, releases its worker by parking its own
queue row like a workflow-as-code parent, and resumes with the answer (approved,
approver, comment) or a timeout once someone answers from the run page.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): tie approval answers to their request and hold the turn

- A turn that asks for approval runs none of its other calls; they are refused
  so the model repeats them after the answer.
- The run page sends the request's call id. An answer naming another request is
  ignored and the agent parks again for the time left; `suspend` is zeroed when
  the answer is read, so answers sent while the agent runs are refused.
- The checkpoint is read whatever the current roster holds, the answer's
  outcome is written by position, and a cancel that beats the park completes
  the job as cancelled.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): say why calls beside an invalid approval request were refused

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): one approval form per request, none on a finished run

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): answer approval requests from a flow's chat

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): show chat approvals for agents that do not stream

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): approval as a per-tool setting, built-in tool asks a question

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): say when a nested agent leaves out tools that require approval

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): show chat approvals and questions on the tool's own card

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): question choices, and approval buttons shared with approval steps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): show a waiting agent as suspended, and refused calls as not run

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): each added chat card shows its own call, answered marker keyed by job

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(ai-agent): approve and reject a held call through resume_suspended

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): require approval in a tool settings tab, plain descriptions on built-in tool panes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): sentence case for the web search tool label

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* feat(ai-agent): built-in tool drawers explain themselves in the header subtitle

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): clearer question tool subtitle

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(ai-agent): badge the tool settings tab like run settings

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* chore(ai-agent): regenerate system prompts and MCP session files

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 19:22:45 +01:00
Ruben FiszelandClaude Opus 5.5 a2b91fd17e fix: detect attached -c and --no-config in DENO_FLAGS (#12027)
* fix: detect attached -c and --no-config in DENO_FLAGS

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: keep tracing span on handle_deno_job

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 19:20:58 +01:00