Repository navigation
docs(cron): accept keeping a server whose start-up sign-in failed - #17121
iamwhatever merged 1 commit into
Conversation
GPT 6.1 Review (fork) — ✅ no blocking findingsReviewed Review detailsNo findings. |
Opus 5.5 Review (fork) — ✅ no blocking findingsReviewed |
Design Review (Opus 5.5, fork) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing to check. [DESIGN-REVIEWED] d4f9794 Description read[DESCRIPTION-READ] 8c70a5cdcd760f4c099619d6a7396b46ed7de2a91343131f3cffc24c2cca93dd 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 |
056affe to
7da8bb1
Compare
7da8bb1 to
c652118
Compare
…accepted A server whose start-up sign-in failed and that reports it as a tool error is kept for the rest of the run since kirodotdev#16905. The spec now says that this is accepted and why: the server's own retry rules decide when it signs in again, the next run starts fresh servers, and restarting it after each error would sign in again at once. Refs kirodotdev#16659
c652118 to
d4f9794
Compare
iamwhatever
left a comment
There was a problem hiding this comment.
Approving (crew-pr-reviewer). One spec sentence group in learn-cron-dashboard.md. It states that keeping a script cron's MCP server after a failed start-up sign-in is accepted, and why. That answers the question the #16905 approval asked. Docs only. CI is green and the base is main.
Problem / Motivation
Goal: The cron spec says whether keeping a script cron's MCP server after a failed start-up sign-in is accepted, as the #16905 approval asked.
Since #16905, a call the MCP server answers with a tool error keeps the run's server. That includes a server whose start-up sign-in failed and that reports the failure as a tool error: it stays for the rest of the run instead of being restarted. The approval on #16905 asked: "say in the spec whether you accept it."
Why it matters
Without the sentence, a reader of the spec cannot tell whether this failure mode is known and chosen, or an oversight.
Not a goal
What changed (motivation → approach → change)
The "Script-mode MCP server lifetime" bullet in
docs/system-specs/modules/learn-cron-dashboard.mdgains three sentences: keeping such a server is accepted. The server's own retry rules decide when it signs in again. A server that never retries fails only until the run ends, because the next run starts fresh servers. Restarting it after each such error would sign in again at once, which is the repeated sign-in that keeping servers avoids.The Slack server behind #16659 recovers within the run: after a failed sign-in it fails calls fast until a cooldown passes (Slack's Retry-After, capped at five minutes), then signs in again on the next call.
Backwards compatibility
Compatible: documentation only, no code changes.
Removes nothing: one spec bullet gains three sentences.
Tests
N/A, documentation only. The behaviour the sentences describe is already pinned by
test_a_tool_error_keeps_its_server_for_the_next_callintest/test_cron_script.py. The docs tests that read this spec pass (480 tests in the six test files that name it).Manual verification
N/A, documentation only: the sentences restate tested behaviour and add no new claim about the code.
Related Issues
Refs #16659
Follow-up to #16905: the Design note its approval asked to answer in the spec.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)