fix(memory): close local embedding providers on timeout by clawsweeper[bot] · Pull Request #84048 · openclaw/openclaw

clawsweeper · 2026-05-19T08:07:52Z

Makes #83858 merge-ready for the ClawSweeper automerge loop.
The edit pass should inspect the live PR diff, review comments, and failing checks; rebase if needed; keep the contributor branch credited; and stop only when validation is green or an external blocker is proven.

ClawSweeper 🐠 replacement reef notes:

Cluster: automerge-openclaw-openclaw-83858
Source PRs: fix(memory): close local embedding providers on timeout #83858
Credit: Source PR: fix(memory): close local embedding providers on timeout #83858
Validation: pnpm check:changed
Replacement reason: ClawSweeper could not update the source PR branch directly, so it opened a writable replacement PR instead.
Automerge requested by: @hxy91819

Repair fallback: GitHub rejected the repair branch push because it updates workflow files and the ClawSweeper app token does not have workflows permission

Inherited issue-closing references from the source PR:
Fixes #83792

Co-author credit kept:

@brokemac79: Co-authored-by: brokemac79 [email protected]

fish notes: model gpt-5.5, reasoning high; reviewed against 70e7256.

clawsweeper · 2026-05-19T08:09:17Z

Codex review: passed.

Workflow note: Future ClawSweeper reviews update this same comment in place.

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

Summary
The branch adds a close lifecycle for local memory embedding providers, scoped memory search/index teardown for one agent, Active Memory timeout cleanup, focused tests, and a changelog entry.

Reproducibility: yes. The linked issue gives a concrete OpenClaw 2026.5.18 Telegram Active Memory timeout path with retained local GGUF mapping, and current-main source inspection confirms there is no timeout cleanup for that local embedding provider path.

PR rating
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
Summary: Strong live proof and focused tests support the patch, while the lifecycle scope and availability tradeoff make this a normal mergeable PR rather than exceptional.

Rank-up moves:

none

What the crustacean ranks mean

🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

PR egg
✨ Hatched: 💎 rare Velvet Proofling

       _..------.._          
    .-'  .-.  .-.  '-.       
   /    ( * )( * )    \      
  |        .--.        |     
  |   <\   ====   />   |     
   \    '.______.'    /      
    '-._   ____   _.-'       
        `-.____.-'           
       __/|_||_|\__          
      /__.'    '.__\         
       `-----------'         
 +===================+

Rarity: 💎 rare.
Trait: watches the merge queue.
Share on X: post this hatch
Copy: My PR egg hatched a 💎 rare Velvet Proofling in ClawSweeper.

What is this egg doing here?

Eggs appear after the PR passes real-behavior proof. It is here for vibes, not verdicts: it does not change labels, ratings, merge decisions, or automation.
The shell reacts to review momentum: open follow-up work warms it up, re-review makes it wobble, and a clean final review lets it hatch.
How to hatch it: reach status: 👀 ready for maintainer look or status: 🚀 automerge armed; that usually means sufficient real-behavior proof, no blocking P0/P1/P2 findings, no security attention needed, and clean correctness.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

Real behavior proof
Sufficient (logs): Sufficient logs: the source PR body includes after-fix live VPS Telegram timeout logs showing the reply completed, GGUF/node-llama map count stayed zero, and RSS returned near baseline.

Risk before merge
Why this matters: - The cleanup intentionally evicts scoped memory-search resources after an Active Memory timeout, so local embedding users may trade retained RSS for a later cold start on the same agent.

The proof is strong for one Linux VPS Telegram timeout path, but it is not a long soak across repeated topics, non-Linux hosts, or every alternate memory provider.

Maintainer options:

Accept Scoped Timeout Cleanup (recommended)
Accept the cold-start tradeoff because it releases retained local embedding mappings on the proven timeout path.
Pause For Broader Cache Policy
Pause only if maintainers want reference-counted or configurable cache eviction before changing timeout cleanup behavior.

Next step before merge
No repair lane is needed because the patch review found no actionable code defect; merge remains gated by normal exact-head checks and maintainer automerge policy.

Security
Cleared: The diff changes memory lifecycle code, tests, and type declarations without adding dependencies, workflow permissions, install hooks, secret handling, or downloaded code execution.

Review details

Best possible solution:

Land the scoped timeout cleanup once required exact-head checks pass, keeping the lifecycle tests and accepting the local-embedding cold-start tradeoff as the bounded fix for retained mappings.

Do we have a high-confidence way to reproduce the issue?

Yes. The linked issue gives a concrete OpenClaw 2026.5.18 Telegram Active Memory timeout path with retained local GGUF mapping, and current-main source inspection confirms there is no timeout cleanup for that local embedding provider path.

Is this the best way to solve the issue?

Yes. The PR fixes the implicated owner boundary by adding a provider close lifecycle and wiring Active Memory timeout cleanup through memory-core rather than adding a new config knob or broad cache policy.

Label justifications:

P2: This is a normal-priority bug fix for retained local embedding memory after Active Memory timeouts with limited affected surface.
merge-risk: 🚨 availability: Merging can intentionally close warm local memory-search resources after timeouts, causing later cold starts or same-agent memory-search disruption.

What I checked:

Current main lacks local provider close: On current main, createLocalEmbeddingProvider caches llama/model/context objects and returns embed methods without any provider close lifecycle, matching the retained-resource failure mode. (packages/memory-host-sdk/src/host/embeddings.ts:33, aef93881af5f)
Current main timeout path does not release memory search resources: The Active Memory timeout branches record the timeout and return without closing the agent's memory search/index manager on current main. (extensions/active-memory/index.ts:2737, aef93881af5f)
PR adds provider abort/dispose lifecycle: The PR head adds closed-state checks, AbortController wiring for resolve/load/context creation, and a close method that aborts initialization and disposes context, model, and llama runtime resources. (packages/memory-host-sdk/src/host/embeddings.ts:57, 8e2e369b5cfd)
PR wires timeout cleanup through Active Memory and memory-core: The PR schedules scoped cleanup after Active Memory timeouts and routes closeMemorySearchManager through QMD and builtin index-manager teardown for the affected agent. (extensions/active-memory/index.ts:995, 8e2e369b5cfd)
Dependency contract checked: node-llama-cpp 3.18.1 type declarations expose loadSignal on LlamaModelOptions, createSignal on LlamaEmbeddingContextOptions, dispose methods on Llama/LlamaModel/LlamaEmbeddingContext, and signal/directory on ResolveModelFileOptions.
Linked issue has a concrete live repro: The linked issue documents OpenClaw 2026.5.18 on a Linux VPS with Telegram Active Memory, local memorySearch provider, a timeout path, and a retained 314 MB embedding GGUF mapping while the gateway was idle.

Likely related people:

steipete: GitHub path history shows repeated recent work on Active Memory routing, memory-host embeddings, and memory-core lifecycle surfaces, including timed-out embedding request handling. (role: recent area contributor; confidence: high; commits: a6225060f1dc, 30e2654ac2c8, a38c2c233a44; files: packages/memory-host-sdk/src/host/embeddings.ts, extensions/active-memory/index.ts, extensions/memory-core/src/memory/manager.ts)
Takhoffman: Recent current-main history shows work on Active Memory custom recall tools, which is adjacent to the timeout and recall-resource path changed here. (role: active-memory feature contributor; confidence: medium; commits: 2f2602508511; files: extensions/active-memory/index.ts)
giodl73-repo: Recent current-main history shows work in memory-core manager behavior around stale session catch-up and indexing, adjacent to the index manager lifecycle touched by this PR. (role: memory-core recent contributor; confidence: medium; commits: 8af2af24a53c; files: extensions/memory-core/src/memory/manager.ts)

Codex review notes: model gpt-5.5, reasoning high; reviewed against aef93881af5f.

clawsweeper · 2026-05-19T08:17:25Z

🦞🧹
ClawSweeper automerge is enabled.

Head: 8e2e369b5cfd
Label: clawsweeper:automerge
Action: exact-head review queued (workflow sweep.yml, event repository_dispatch).
Flow: review this head, repair/rebase only if needed, then re-review the exact repaired head before merge.

Draft PRs stay fix-only until GitHub marks them ready for review. Pause with /clawsweeper stop.

Automerge progress:

2026-05-19 08:43:28 UTC repair completed 70e725614006 (no branch change) in 21m 18s Run: https://github.com/openclaw/clawsweeper/actions/runs/26085078946 Codex fix worker timed out after 1200000ms

2026-05-19 08:43:27 UTC review queued 70e725614006 (after no-op repair)

2026-05-19 08:43:30 UTC repair finished (blocked) in 21m 20s Run: https://github.com/openclaw/clawsweeper/actions/runs/26085078946 Codex fix worker timed out after 1200000ms

2026-05-19 08:52:27 UTC repair queued 70e725614006 (autonomous) Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454

2026-05-19 08:57:29 UTC repair started (running) in 1s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 automerge-openclaw-openclaw-83858

2026-05-19 08:57:49 UTC validation plan (passed) in 21s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 pnpm check:changed; node scripts/run-vitest.mjs packages/memory-host-sdk/src/host/embeddings.test.ts; node scripts/run-vitest.mjs extensions/active-memory/inde...

2026-05-19 08:58:13 UTC Codex write preflight (passed) in 44s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 danger-full-access

2026-05-19 09:05:03 UTC Codex edit 1 ac5dd87c4454 (complete) in 7m 34s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 exit 0

2026-05-19 09:11:43 UTC validation and review 1 8e2e369b5cfd (complete) in 14m 15s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 already-current

2026-05-19 09:11:49 UTC repair completed 8e2e369b5cfd (branch updated) in 14m 21s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 initial automerge rebase is delegated to Codex repair

2026-05-19 09:11:49 UTC review queued 8e2e369b5cfd (after repair)

2026-05-19 09:19:05 UTC automerge wait 8e2e369b5cfd (ready) in 21m 37s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 checks and exact-head review are ready

2026-05-19 09:19:07 UTC repair finished 8e2e369b5cfd (pushed) in 21m 38s Run: https://github.com/openclaw/clawsweeper/actions/runs/26086777454 repair_contributor_branch

2026-05-19 09:18:55 UTC review passed 8e2e369b5cfd (structured ClawSweeper verdict: pass (sha=8e2e369b5cfd8081876b2b2b5d86b2c1cc603...)

2026-05-19 09:19:13 UTC merged 8e2e369b5cfd (merged by ClawSweeper automerge)

2026-05-19 09:19:18 UTC review queued 8e2e369b5cfd (queued)

hxy91819 · 2026-05-19T08:22:15Z

/clawsweeper automerge with thanks to brokemac79

clawsweeper · 2026-05-19T08:24:10Z

🦞👀
ClawSweeper is here and listening for maintainer commands.

Supported commands: /review, /clawsweeper status, /clawsweeper re-review, /clawsweeper implement, @clawsweeper fix, /clawsweeper build, /clawsweeper fix ci, /clawsweeper address review, /clawsweeper rebase, /clawsweeper autofix, /clawsweeper automerge, /clawsweeper approve, /autoclose <reason>, /clawsweeper explain, /clawsweeper stop.

I only act for maintainers, or for trusted ClawSweeper feedback on a ClawSweeper PR or PR opted into clawsweeper:autofix or clawsweeper:automerge.

hxy91819 · 2026-05-19T08:25:34Z

/clawsweeper automerge

Summary: - The branch adds a close lifecycle for local memory embedding providers, scoped memory search/index teardown for one agent, Active Memory timeout cleanup, focused tests, and a changelog entry. - Reproducibility: yes. The linked issue gives a concrete OpenClaw 2026.5.18 Telegram Active Memory timeout pa ... current-main source inspection confirms there is no timeout cleanup for that local embedding provider path. Automerge notes: - PR branch already contained follow-up commit before automerge: fix(memory): close local embedding providers on timeout Validation: - ClawSweeper review passed for head 8e2e369b5cfd8081876b2b2b5d86b2c1cc6036bf. - Required merge gates passed before the squash merge. Prepared head SHA: 8e2e369b5cfd8081876b2b2b5d86b2c1cc6036bf Review: openclaw/openclaw#84048 (comment) Co-authored-by: brokemac79 <[email protected]> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: hxy91819 Co-authored-by: hxy91819 <[email protected]>

Summary: - The branch adds a close lifecycle for local memory embedding providers, scoped memory search/index teardown for one agent, Active Memory timeout cleanup, focused tests, and a changelog entry. - Reproducibility: yes. The linked issue gives a concrete OpenClaw 2026.5.18 Telegram Active Memory timeout pa ... current-main source inspection confirms there is no timeout cleanup for that local embedding provider path. Automerge notes: - PR branch already contained follow-up commit before automerge: fix(memory): close local embedding providers on timeout Validation: - ClawSweeper review passed for head 8e2e369. - Required merge gates passed before the squash merge. Prepared head SHA: 8e2e369 Review: openclaw#84048 (comment) Co-authored-by: brokemac79 <[email protected]> Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com> Co-authored-by: clawsweeper[bot] <274271284+clawsweeper[bot]@users.noreply.github.com> Approved-by: hxy91819 Co-authored-by: hxy91819 <[email protected]>

…ctive-memory recall timeout Active-memory recall-timeout cleanup called closeActiveMemorySearchManager, which retired the agent-scoped shared QMD manager used by chat memory_search and startup timers/watchers. A request-scoped recall timeout then left chat memory_search returning "memory search manager is closed" (and entering its 60s cooldown) even while QMD and CLI search stayed healthy. The cleanup's legitimate job (openclaw#84048) is to release the disposable local builtin index/embedding providers a hung recall can leave open, not to retire the shared manager. Add a narrow releaseMemoryIndexManagers path, exposed through the memory-host-search SDK facade as releaseActiveMemoryIndexManagers, that releases only those local index managers and leaves the shared QMD manager alive, and point active-memory's timeout cleanup at it. closeMemorySearchManager keeps its full-close contract for callers that intentionally retire an agent's manager. The recall's in-flight QMD query is already cancelled via the embedded run abort signal. Fixes openclaw#96455

…ctive-memory recall timeout Active-memory recall-timeout cleanup called closeActiveMemorySearchManager, which retired the agent-scoped shared QMD manager used by chat memory_search and startup timers/watchers. A request-scoped recall timeout then left chat memory_search returning "memory search manager is closed" (and entering its 60s cooldown) even while QMD and CLI search stayed healthy. The cleanup's legitimate job (openclaw#84048) is to release the disposable local builtin index/embedding providers a hung recall can leave open, not to retire the shared manager. Add a request-scoped scope: "index-managers" mode to closeActiveMemorySearchManager (routing to a narrow releaseMemoryIndexManagers path in memory-core) that releases only those local index managers and keeps the shared QMD manager alive, and point active-memory's timeout cleanup at it. The default full-close behavior is unchanged for callers that intentionally retire an agent's manager, so the shipped SDK export surface does not grow. The recall's in-flight QMD query is already cancelled via the embedded run abort signal. Fixes openclaw#96455

clawsweeper Bot mentioned this pull request May 19, 2026

fix(memory): close local embedding providers on timeout #83858

Closed

openclaw-barnacle Bot removed the proof: supplied External PR includes structured after-fix real behavior proof. label May 19, 2026

brokemac79 and others added 3 commits May 19, 2026 08:59

fix(memory): close local embedding providers on timeout

b000829

fix(memory): close local embedding providers on timeout

ac5dd87

fix(memory): close local embedding providers on timeout

8e2e369

clawsweeper Bot force-pushed the clawsweeper/automerge-openclaw-openclaw-83858 branch from 70e7256 to 8e2e369 Compare May 19, 2026 09:11

clawsweeper Bot merged commit 1c1c75d into main May 19, 2026
98 checks passed

clawsweeper Bot deleted the clawsweeper/automerge-openclaw-openclaw-83858 branch May 19, 2026 09:19

clawsweeper Bot mentioned this pull request May 25, 2026

Fix local embedding worker safety #85348

Merged

ZengWen-DT mentioned this pull request Jun 24, 2026

fix(memory): keep shared QMD manager alive on active-memory recall timeout #96459

Closed

clawsweeper Bot mentioned this pull request Jun 24, 2026

fix(memory): separate active-memory resource release from manager close #96465

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(memory): close local embedding providers on timeout#84048

fix(memory): close local embedding providers on timeout#84048
clawsweeper[bot] merged 3 commits into
mainfrom
clawsweeper/automerge-openclaw-openclaw-83858

clawsweeper Bot commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 19, 2026 •

edited

Loading

Uh oh!

hxy91819 commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026

Uh oh!

hxy91819 commented May 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

clawsweeper Bot commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

clawsweeper Bot commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

hxy91819 commented May 19, 2026

Uh oh!

clawsweeper Bot commented May 19, 2026

Uh oh!

hxy91819 commented May 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

clawsweeper Bot commented May 19, 2026 •

edited

Loading

clawsweeper Bot commented May 19, 2026 •

edited

Loading