Skip to content

feat(tracing): worker creation, report-to-wake links, manager changes and patrol counts (5b, #1643) - #37088

Open
kyleseaman wants to merge 1 commit into
mainfrom
crew/devmate/otel-pr5b
Open

kyleseaman wants to merge 1 commit into
mainfrom
crew/devmate/otel-pr5b

Conversation

@kyleseaman

@kyleseaman kyleseaman commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: With tracing on, a conductor mode or Bot conductor's trace shows who started each worker, which reports each wake answered, when a report never arrived, who manages each session, and the board's health on every patrol.

5a made a round one trace, but the edges between sessions are still missing. A worker's first turn does not point back at the conductor turn that created it. A wake turn says nothing about which reports it is answering, so a batch of three finished workers looks like an unexplained conductor turn. A report the harness never delivered leaves no trace at all. A handoff, adopt or release changes who reads a session, and the trace cannot show it. And the stale, orphaned and overdue counts the conductor acts on live only in its tool output.

Why it matters

These are the questions an operator asks of a conductor mode run when something goes wrong: why did this worker exist, why did the conductor wake, why did it not hear from item X, and who was supposed to be watching. The tracing RFC (§5.5, phase 5) lists each of them. Without them a round's waterfall shows the work but not the causes.

Not a goal

What changed (motivation → approach → change)

Design record: the native tracing RFC §5.5 and §10 row 5; this is 5b, the half split off from #36833.

Approach: each producer already has one place where the fact is known, so each gets one hook there, and the turn that answers a fact gets a link when it binds.

Span or attribute Where What it says
kirocrew.session.create / .fork the session_create / session_fork routes (work_dispatch creates through the same route) in the calling conductor turn; the new session's first traced turn links to it once
one per report on a patrol turn nudge_dashboard_fire.py marks the turn a work-ledger watch delivers (work.patrol_turn); dashboard/trace_binding.py adds the links when the turn binds, and chat_turn/queue_cycle.py spends an unclaimed mark at the turn's queue hand-off each kirocrew.work.report and undelivered event written on the board since the last patrol turn, at most 64
kirocrew.work.report_undelivered src/kiro_crew/undelivered_call_watch.py's record_undelivered_call, next to its wake a span under the item: the worker tried to report and the call never arrived
kirocrew.session.manager_change the handoff route and session_handoff.hand_back_on_close old and new manager, hashed; in the calling turn, or alone for a close's hand-back
kirocrew.session.parent_change the adopt and release routes old and new tree parent, hashed (adopt and release move the tree edge, not the manager); in the calling turn
kirocrew.work.{open,stale,orphaned,pending_overdue,undelivered}_items the conductor's own whole-board work_ledger_read counts on the reading turn's span, from the read's own flags
  • New module tracing/workers.py (creation, manager changes and tree-parent changes); the rest is in tracing/work.py. Named apart from feat(tracing): session lifecycle spans: open, close, reset, compaction, switch, adopt (6, #1643) #36830's tracing/sessions.py so the two do not collide.
  • Eleven new allowlisted keys: kirocrew.session.op and the five hashed session keys (target, manager, previous manager, parent, previous parent), plus five counts. Values are ids, a closed set of operations, and counts: never a title or any text.
  • Spec: metrics.md Traces, in this commit.

Rejected: linking every conductor turn to the pending reports. A person's own turn does not answer them, so only the watch's turns take them; a wake steered into a turn already running adds none, because a span's links are fixed at its start, and the reports wait for the next patrol turn.

Overlap with #36830

This PR's kirocrew.session.parent_change span for an adopt sits in the acting turn's trace and records who moved the session. #36830's adopt event sits in the session's own lifecycle trace and records what happened to the session. So one adopt gives one span in each trace. That is by design, not a duplicate. Merge order: this PR first, then #36830 rebases onto it.

Backwards compatibility

Compatible: tracing is off by default, and off, each hook is one peek() returning None. No route, ledger or wake behaviour changes; the hooks run after the operation committed and never raise.

Removes nothing: the diff adds spans, links and allowlisted attributes.

Tests

test/tracing/test_tracing_work.py gains twenty-two tests, driving the real ledger routes and trace_binding.bind_turn:

  • A worker created in a conductor turn: the create span sits in that turn with the target's hash, the worker's first turn links to it, and its second turn does not.
  • A fork outside any traced turn, or an unknown operation, records nothing.
  • Two workers report; a person's conductor turn takes no links, the next patrol turn links to both report spans, and the patrol turn after it to none.
  • An undelivered report is a span under its item, and the next patrol turn links to it.
  • A board read puts the five counts on the reading turn.
  • A handoff sits in the calling turn with both managers hashed; a hand-back is a trace of its own in the target's session; an unknown operation records nothing. Two moves of one session in one turn stay two spans, keyed by their ordinal in the turn, and the same turn replayed after a restart gives the same two ids; a replayed hand-back keeps its id. A long turn keeps its move count however many other turns move meanwhile, since the count lives on the turn's own trace object.
  • session_create registers the creation link before its first message starts the worker, so the worker's first turn takes it.
  • A cancelled undelivered-report trace still wakes the conductor.
  • The adopt, handoff and release routes pass the right result keys as previous and new manager or parent; an adopt or release is a kirocrew.session.parent_change span with the parent hashes and no manager keys.
  • The dashboard fire runs a work-ledger watch's turn as a patrol turn, and any other loop's turn not. The patrol mark is one-shot: a turn fair admission runs in a new task still takes the links, a queued successor the patrol turn starts after it bound takes none, and a patrol turn that never bound (refused at admission, or tracing off) spends the mark at its queue hand-off, so its successor takes none either.
  • A bind cancelled during its crew read (a tab closed) leaves the worker's creation link and the board's report links pending for the next bind; a crew read that merely fails still binds and takes them.
  • Off: every hook records and keeps nothing.

test_session_handoff.py: a close's hand-back records one manager change per session moved, naming the creator as the new manager when the move removes an override. test_tracing_ids_attrs.py: the pinned allowlist gains the eleven keys.

Determinism: 20 runs at -n0 and 10 at -n 4 --dist loadgroup --max-worker-restart=2 beside test_tracing_runs.py, test_session_handoff.py and test_chat_runner_coverage_queue.py, all green; a shuffled order (seed 4219) green. No sleeps.

Local: related backend tests, including the chat queue hand-off and workflow suites (12,406 passed; the 224 red ones, 221 in the test_ai_review_workflows_* CI-script tests and 3 in test_design_tweak_backend_queue.py, fail the same on a clean origin/main worktree on this host; an earlier wider run's 10 app-backend, app-scaffold, codex-ACP and client-binding reds did too), semgrep rules on the diff (0 findings), the flake-pattern ratchet and the pre-push gates pass. Two test_tracing_otlp.py tests (test_the_exit_flush_sends_spans_still_queued, test_a_traced_turn_reaches_a_loopback_receiver_with_its_vault_header) fail on origin/main itself, from phase 4a and #36721; a separate fix PR covers them.

Manual verification

N/A — unit coverage drives the real routes, the real turn binding and the real dashboard fire; no UI or live collector is involved.

Follow-ups

  • A goal's last round still gets no ultra outcome marker: ending it needs a ledger signal for a finished goal.
  • A wake steered into a running turn adds no links; its reports are linked by the next patrol turn.
  • A worker report's link is queued right after the board copy lands, before the worker's mirror, but the board copy's eager publication (crew_log/emit_surfaces._note_eager) can still start a wake that binds during the trace call's thread hop; that wake misses the link and the next patrol turn takes it. Closing it needs a post-append, pre-publication hook in emit_surfaces, which changes the shared crew-log write path, so it is left for its own change.
  • A create made outside a traced turn records no span, and pending first-turn links are lost on a restart.
  • A top turn that binds and then fails before its ACP stream starts (for example AcpAuthRequired while the provider initialises) has already taken its creation or report links, and releasing the unused root does not give them back; restoring them on release, or taking them when TurnTrace starts, would close it.
  • A manager reopened from History that closes again with the same hand-back gets the same span id as its first close, so a backend keeps one of the two; a per-close stamp from the close record would separate them.
  • The queue hand-off spends an unclaimed patrol mark after the turn tail's first awaits; spending it at tail entry, or when a top turn binds, would also cover a task the turn started that binds during those awaits (none exists today).
  • The manager-move count is an attribute set on TurnTrace rather than declared in its __init__; declaring it would make a later __slots__ change fail loudly instead of through the span tests.
  • Route-level tests with tracing on: the whole-board guard in work_ledger_read.py, the fork hook, and the key pairs session_control.py passes for adopt, handoff and release are covered only through the helpers today.
  • Tests for the caps: at most 64 answered links per patrol turn (MAX_ANSWERED_LINKS) and at most 4096 boards (_MAX_BOARDS).
  • Remove work.in_patrol_turn(): it has no production caller, and the tests could read _PATROL directly.

Related Issues

Part of #1643

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@kyleseaman
kyleseaman requested review from a team and KiroCrewCaptain as code owners November 14, 2026 04:55
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 14, 2026
@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 4643b77751de24a4626b2972ee7082f6803a25e5 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 4643b77751de24a4626b2972ee7082f6803a25e5 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: all of it is additive tracing that stays off by default, each fact is hooked where it is already known, and nothing is taken away.

Notes

  • work_ledger_worker.api_work_report now awaits trace_work.recorded before the mirror append. The window this leaves (a wake from eager publication that misses the link) is written down as a shortcut:; the post-append hook in emit_surfaces that closes it should get a tracking issue so it doesn't stay a comment.

[DESIGN-REVIEWED] 4643b77

Description read

[DESCRIPTION-READ] 3506f2297ee5cc077c464e4046b8f7dc4340a15eca0508d74ebc19f510ec576e

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 4643b77751de24a4626b2972ee7082f6803a25e5 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

Every item maps to a §5.5 producer in the tracing RFC (base status partial), and tracing is off by default. The one dead surface is a test-only public helper.

Notes

  • Item 9 — zero consumers: def in_patrol_turn() -> bool: in tracing/work.py. Grepping in_patrol_turn under src/ finds only its definition, so it has 0 production consumers; its 5 uses are all in test/tracing/test_tracing_work.py. Subtraction: delete in_patrol_turn, and have the tests check _PATROL.get().taken directly.

What this change ships

Inventory (9 items) — 8 justified

Intent: let an operator see in a conductor's trace who started each worker, which reports each wake answered, which reports never arrived, who manages each session, and the board's health. This is an ADDITION.

  1. A worker's create/fork shows up as a span in the conductor turn, and the worker's first turn links back to it — justified
  2. A work-ledger watch's patrol turn links to each report since the last patrol, at most 64 — justified
  3. An undelivered report becomes a span under its item; the wake still fires after the trace — justified
  4. A handoff or close hand-back becomes a manager-change span with hashed managers — justified
  5. An adopt or release becomes a parent-change span with hashed parents — justified
  6. A whole-board ledger read puts five health counts on the reading turn — justified
  7. Eleven new keys in the attribute allowlist — justified
  8. The work-report route refuses an unrecorded report earlier and queues the trace link before the mirror append — undeclared
  9. A new public in_patrol_turn() that only tests call — zero consumers

[FIRST-PRINCIPLES-REVIEWED] 4643b77

Description read

[DESCRIPTION-READ] 4b27ce172e96b43841171003e523b1bc7b76aa090a8a250ad572157aba6ab270

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@KiroCrewCaptain

KiroCrewCaptain commented Nov 14, 2026 •

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: A person should decide

Why: the decider estimates a 12% chance that this change introduces a defect, within the 12% limit for approving on its own, but it changes 1272 lines, more than the 400 it may approve on its own.

Additive, off-by-default tracing hooks for worker creation, report-to-patrol links, undelivered reports, manager and parent changes, and board-health counts. Checked against the code: pending report links are stored and looked up under the same board key. The patrol mark is set only for the work-ledger watch's turn, can be claimed once, carries into a new task, and is spent at the queue hand-off. Moving the work_report unlanded-refusal earlier keeps behaviour the same, because the mirror was already skipped when the record didn't land. The undelivered wake still fires in a finally if the trace is cancelled. Create/fork/handoff/adopt/release result keys match what the routes return, and the hand-back loop handles both shapes of the change map. Every hook catches its own errors. Allowlist and spec are updated, and the tests drive the real routes and binding. The CI reds are redaction tests that don't exist at this head, so they come from the newer main CI merged with, not from this PR.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit 4643b77751 · 1272 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 4643b77751de24a4626b2972ee7082f6803a25e5 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 4643b77

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 4643b77751de24a4626b2972ee7082f6803a25e5 false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 14, 2026
@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from 1ca0e7b to 5ba5c3d Compare November 14, 2026 06:17
@kyleseaman

kyleseaman commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor Author
  • span=9be2862bbfcf session_handoff.py:259: record the creator who regains management on hand-back

fixed in 5ba5c3d: when a close removes a session's manager override, the hand-back span now names the open session's creator (the manager session_managers.holder falls back to) instead of an empty manager, so kirocrew.session.manager_hash is present; a closed worker still records none.
test_a_hand_back_is_one_manager_change_span_per_session now covers the hand-up, the creator return and a closed worker against real spans; the creator-return case fails on 1ca0e7b with KeyError: 'kirocrew.session.manager_hash'.

@github-actions

github-actions Bot commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 4643b77751de24a4626b2972ee7082f6803a25e5 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 4643b77

Verdict parsed from the review's SHA-scoped output markers for commit 4643b77751de24a4626b2972ee7082f6803a25e5.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 4643b77751de24a4626b2972ee7082f6803a25e5 false-positive: <reason, 20 to 500 characters>

@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from 5ba5c3d to fbb101f Compare November 14, 2026 07:21
@kyleseaman

Copy link
Copy Markdown
Contributor Author
  • span=4b74fd25fd8c session_control.py:1225: tree changes export false management transfers

fixed in fbb101f: adopt and release move only the tree edge (session_tree.adopt_target / release_target never write session_managers), so they are now a separate kirocrew.session.parent_change span carrying kirocrew.session.parent_hash / previous_parent_hash and no manager keys; kirocrew.session.manager_change is kept for handoff and hand-back. Ids use session.parent_change as their first part and keep the per-turn ordinal; metrics.md and the allowlist pin are updated.
New tests assert both adopt and release record the parent pair and no manager hashes; they fail on 5ba5c3d.

@github-actions github-actions Bot added readiness: checking Automated validation is still running merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Nov 14, 2026
@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from fbb101f to 9dec790 Compare November 14, 2026 08:50
@kyleseaman

Copy link
Copy Markdown
Contributor Author

Rebased: main moved in #36481 (host memory pressure on the work ledger) and #37005 (OTel steps inside a turn), which conflicted with this PR. Head fbb101f -> 9dec790 on main 12a03ac.

No behaviour change in this PR's own code. Related tests (7,036), the pre-push gates and semgrep pass on the new head.

@github-actions github-actions Bot added merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 14, 2026
@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from 9dec790 to 5e2c63a Compare November 14, 2026 09:49
@kyleseaman

Copy link
Copy Markdown
Contributor Author

Rebased: main moved in #37095 (finish the conductor mode rename), which conflicted with this PR in tracing/work.py. Head 9dec790 -> 5e2c63a on main 16e11a1.

  • tracing/work.py: kept this PR's patrol-mark block (_answerable, MAX_ANSWERED_LINKS, _Patrol, _PATROL) and main's KIND_ROUND = "conductor_mode", without the old "stays ultra" comment.
  • None of this PR's added lines use a name refactor: finish the conductor mode rename #37095 renamed.

No behaviour change in this PR's own code (range-diff checked by an independent verifier). Related tests (7,036), the pre-push gates and semgrep pass on the new head.

@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 14, 2026
@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from 5e2c63a to c7d358b Compare November 14, 2026 11:00
@kyleseaman

Copy link
Copy Markdown
Contributor Author
  • span=b66c053ac95c tracing/work.py:438: report links register after their wake can already bind

fixed in c7d358b (partly): the report route now runs trace_work.recorded right after the authoritative record lands and before the worker's mirror append, so a slow mirror (the stated trigger) can no longer delay the link past the wake; a report that does not land is still refused before any trace. test_the_report_link_is_ready_when_the_workers_mirror_lands drives the real route and asserts the patrol link is already there when the mirror is written; it fails on 5e2c63a.
Residual race, accepted: the board copy's eager publication happens inside the crew-log append itself (crew_log/emit_surfaces._note_eager), so a wake that dispatches, passes admission and binds within the trace call's single thread hop still misses the link, which the next patrol turn then takes. Closing that needs a post-append, pre-publication hook in emit_surfaces, which changes the shared crew-log write path for a trace-attribution gap; it is listed in the PR body's follow-ups and noted with a shortcut: comment at the call.

@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from c7d358b to c801216 Compare November 14, 2026 11:57
@kyleseaman

Copy link
Copy Markdown
Contributor Author
  • span=0b9596b98dee dashboard/trace_binding.py:97: cancelled binding permanently consumes causal links

fixed in c801216: bind_turn now takes the worker's creation link (workers.first_turn_link) and the board's report links (work.answered_links) after the awaited crew read, immediately before bind_turn_root with no await between, so a cancellation during the read (a tab closed) leaves both pending; the unclaimed patrol mark is then spent by the queue hand-off, and the next patrol turn takes the reports. An ordinary read failure still binds and takes them, and the takes keep their own exception guard so the function still never raises.
New tests cancel a worker's first bind and a patrol bind inside _crew_attrs and assert the next bind still links to the creation span and the report span; both fail on c7d358b. A third pins that a failed (not cancelled) read still consumes the links.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Nov 14, 2026
@kyleseaman

Copy link
Copy Markdown
Contributor Author
  • span=0b9596b98dee dashboard/trace_binding.py:114: preserve causal links until an ACP turn actually starts

Rebutted as disproportionate (the adjudication downgraded it on the same grounds): the gap is real but narrow. A top turn that binds and then fails before its ACP stream starts (AcpAuthRequired while the provider initialises) has taken its creation or report links, and release_turn_root does not put them back, so that worker's first traced turn after sign-in, or the next patrol turn, lacks the link. Tracing is off by default, the loss is one link on one turn, and no span or ledger record is wrong. Fixing it means either putting links back on release (a restore path in tracing/turns.py, which phase 4a owns) or deferring the take until TurnTrace starts (moving the take out of bind_turn into the streaming seam). Both widen this PR past its scope.
Listed in the PR body's follow-ups: give back session-scoped links when a root is released with no ACP attempt.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Nov 14, 2026
@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

This changes behaviour users already rely on, so it needs a design sign-off from a maintainer before it can merge. A maintainer will post the decision here.

To propose a design, follow the request-for-change process in CONTRIBUTING.md (docs/request-for-change/).

Review result on this head: nothing blocks. We read the whole diff. The work-report reorder changes no behaviour, because the mirror append was already gated on landed. The undelivered-call trace sits in a try/finally, so the conductor wake still fires on cancel. The link keys match on both sides (create and fork targets, conductor ledger keys, and the adopt, handoff and release result keys). Every new attribute goes through the allowlist, session keys appear only as keyed hashes, and no title or text reaches a span. The tests drive the real _hand_off_queue, bind_turn and hand_back_on_close seams instead of mocks, and PR Readiness is green.

Optional:

  • Both this PR and feat(tracing): session lifecycle spans: open, close, reset, compaction, switch, adopt (6, #1643) #36830 trace adopt (here as kirocrew.session.parent_change, there inside adopt_target and _move_tree_edge). Once both land, one adopt or handoff emits two tree-edge spans. Naming one owner and the merge order in both bodies would avoid that.
  • The route-level hooks (the whole-board guard in work_ledger_read.py, the fork hook, and the key pairs passed for adopt, handoff and release in session_control.py) are only tested through the helpers. One test per route with tracing on would pin them. The 64-link and 4096-board caps also have no test.
  • in_patrol_turn() in tracing/work.py has no production caller. The tests could read _PATROL directly, and the function could go.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: passed Eligible automated validation passed for the current revision readiness: checking Automated validation is still running labels Nov 14, 2026
@bolichen97

Copy link
Copy Markdown
Contributor

This head conflicts with main (GitHub reports mergeable: CONFLICTING). Could you rebase onto main? I'll review the full diff on the rebased head, since the conflict may move lines in tracing/work.py or the handlers this PR touches.

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 14, 2026
@kyleseaman
kyleseaman force-pushed the crew/devmate/otel-pr5b branch from c801216 to 4643b77 Compare November 14, 2026 23:49
@kyleseaman

Copy link
Copy Markdown
Contributor Author

Rebased: main moved in #37158 (queue wait and the links a send, a broadcast and a workflow launch leave), which conflicted with this PR. Head c801216 -> 4643b77 on main 5c3b2cd.

  • dashboard/trace_binding.py bind_turn: kept main's links = tuple(links) + sent_links, then this PR's top-turn block that takes the creation and report links. Final link order: sub-agent/workflow, sent, first-turn, answered. No await between the takes and bind_turn_root.
  • tracing/correlation.py and its pinned allowlist test: the union of feat(tracing): queue wait and the links a send, a broadcast and a workflow launch leave (#1643) #37158's kirocrew.queue.wait_ms / kirocrew.launched_workflow_run_id and this PR's work-count and session keys.

No behaviour change in this PR's own code; an independent verifier checked the range-diff and found it unchanged. Related tests (7,117, OTLP extra installed), the pre-push gates and semgrep pass on the new head.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Nov 14, 2026

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: action required A blocking check or review needs attention

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants