fix(app): stop status timeouts from disabling serve and killing waiting one-shots - #1644
Merged
Merged
Conversation
The desktop serve client counted every death, watchdog kills included, against a 3-death budget that never reset, so three sporadic kills over an app run disabled serve for good. The budget now resets on any healthy answer: three generations in a row that never answer still disable it. One-shot CLI runs only heartbeated inside a parse, so a run waiting on another process's cold hydration lock (up to 10 min) went silent and was killed at 45s on every poll. Start the keepalive at command entry.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Telemetry shows
status,yieldandacttimeouts on 0.9.25 desktop installs (61, 48 and 39 installs in 7 days). Two causes:Serve disabled for the whole app run. A serve request that times out kills the child and counts as a death.
deathsnever reset, so three timeouts spread over days of uptime disabled serve for the rest of the run and every panel read fell back to a one-shot. Any healthy answer now resets the count, so the budget is three consecutive deaths with no answer between them (what the existing test name already says). The mac menubar already resets its budget after 300s.One-shots killed while waiting, not while working. The 45s limit is a no-output watchdog.
runParsebeats every 10s, but a one-shot waiting on another process's cold-hydration lock (up to 10 minutes) or aggregating after the parse writes nothing and dies at 45s. One-shots now start the progress keepalive at command entry. It only writes whenCODEBURN_PROGRESS=1, which the desktop app, menubar and tray set; serve keeps its per-request beat. One-shots stay bounded by the 15-minute ceiling in every client.Tests: a resident that answers between watchdog kills keeps restarting (fails on main: capped at 3 starts), and a one-shot blocked on a held hydration lock emits a keepalive (fails on main after 20s of silence). tsc clean at the root and in app/.