(MOT-3748) fix(harness): preserve registrations across engine reloads - #666
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
📝 WalkthroughWalkthroughThe PR adds reusable Harness quickstart workflow inputs and secrets, connects eligible releases to Harness smoke testing, updates Harness dependencies, and adds queue provisioning fallback for engines that reject restart-redelivery configuration. ChangesHarness release and runtime compatibility
Estimated code review effort: 3 (Moderate) | ~25 minutes Sequence Diagram(s)sequenceDiagram
participant ReleaseSetup
participant RegistryPublication
participant HarnessQuickstart
ReleaseSetup->>ReleaseSetup: Detect Harness worker or dependency
ReleaseSetup->>RegistryPublication: Publish release
RegistryPublication->>HarnessQuickstart: Invoke quickstart with channel
Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
skill-check — worker0 verified, 51 skipped (no docs/).
Four for four. Nicely done. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/workflows/release.yml:
- Around line 283-286: Update the reusable workflow invocation of
harness-quickstart.yml to replace secrets: inherit with an explicit secrets
mapping containing only SLACK_BOT_TOKEN and ZAI_API_KEY, preserving the existing
channel input.
In `@harness/src/queue.rs`:
- Around line 46-55: Update ensure_turn_queue around the bounded primary retry
loop so a legacy schema fallback triggered by is_legacy_queue_schema_error can
perform its own retry even when the primary DEFINE_ATTEMPTS budget is exhausted.
Ensure legacy_turn_queue_definition() is actually submitted before returning the
error, and add a regression test covering the schema error on the final primary
attempt and successful fallback provisioning.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 4fa84a8e-eb0b-41ef-96d6-64bd116f0b52
⛔ Files ignored due to path filters (1)
harness/Cargo.lockis excluded by!**/*.lock
📒 Files selected for processing (8)
.github/workflows/harness-quickstart.yml.github/workflows/release.ymlharness/Cargo.tomlharness/iii.worker.yamlharness/src/queue.rsharness/tests/e2e/Cargo.tomlharness/tests/integration/Cargo.tomlharness/tests/manifest.rs
| uses: ./.github/workflows/harness-quickstart.yml | ||
| with: | ||
| channel: ${{ needs.setup.outputs.registry_tag }} | ||
| secrets: inherit |
There was a problem hiding this comment.
🔒 Security & Privacy | 🟠 Major | ⚡ Quick win
Pass only the declared secrets to the reusable workflow.
secrets: inherit exposes every secret available to the release job. The quickstart contract requires only SLACK_BOT_TOKEN and ZAI_API_KEY. Pass those names explicitly.
Proposed fix
with:
channel: ${{ needs.setup.outputs.registry_tag }}
- secrets: inherit
+ secrets:
+ SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
+ ZAI_API_KEY: ${{ secrets.ZAI_API_KEY }}📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| uses: ./.github/workflows/harness-quickstart.yml | |
| with: | |
| channel: ${{ needs.setup.outputs.registry_tag }} | |
| secrets: inherit | |
| uses: ./.github/workflows/harness-quickstart.yml | |
| with: | |
| channel: ${{ needs.setup.outputs.registry_tag }} | |
| secrets: | |
| SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }} | |
| ZAI_API_KEY: ${{ secrets.ZAI_API_KEY }} |
🧰 Tools
🪛 zizmor (1.28.0)
[warning] 283-283: secrets unconditionally inherited by called workflow (secrets-inherit): this reusable workflow
(secrets-inherit)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In @.github/workflows/release.yml around lines 283 - 286, Update the reusable
workflow invocation of harness-quickstart.yml to replace secrets: inherit with
an explicit secrets mapping containing only SLACK_BOT_TOKEN and ZAI_API_KEY,
preserving the existing channel input.
Source: Linters/SAST tools
| if restart_redelivery_requested && is_legacy_queue_schema_error(&last_error) { | ||
| tracing::warn!( | ||
| queue = TURN_QUEUE, | ||
| error = %last_error, | ||
| "queue worker does not support restart redelivery; retrying with the legacy queue schema" | ||
| ); | ||
| payload = legacy_turn_queue_definition(); | ||
| restart_redelivery_requested = false; | ||
| continue; | ||
| } |
There was a problem hiding this comment.
🩺 Stability & Availability | 🟠 Major | ⚡ Quick win
Allow the fallback to consume a retry after the primary budget is exhausted.
If the unsupported-field error occurs on the final DEFINE_ATTEMPTS iteration, this branch changes payload and executes continue, but the bounded loop has no next iteration. ensure_turn_queue then returns the schema error without sending legacy_turn_queue_definition().
The supplied harness/src/main.rs flow waits for queue provisioning before it announces readiness. Give the legacy fallback its own retry budget, and add a regression test that returns the schema error on the final primary attempt.
Also applies to: 132-162
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@harness/src/queue.rs` around lines 46 - 55, Update ensure_turn_queue around
the bounded primary retry loop so a legacy schema fallback triggered by
is_legacy_queue_schema_error can perform its own retry even when the primary
DEFINE_ATTEMPTS budget is exhausted. Ensure legacy_turn_queue_definition() is
actually submitted before returning the error, and add a regression test
covering the schema error on the final primary attempt and successful fallback
provisioning.
1 participant
Published Harness workers can lose their registered functions when the engine reloads another worker, causing calls to fail with
function_not_found. The same release also had a startup incompatibility: its queue recovery option was sent to older published queue workers, so Harness exited during installation.Technical details
iii-sdk/iii-helpers0.21.8release.queue ^0.21.2.redeliver_on_engine_restartwhen an older queue worker rejects that field, preserving startup compatibility while the queue upgrade is rolled out.1.6.7.Fixes MOT-3748
Summary by CodeRabbit
New Features
Bug Fixes
Chores