feat(perf): admit bounded longer deadlines with precise errors - #20
Merged
Conversation
plotarmordev
changed the base branch from
feat/explicit-thinking-11
to
main
September 10, 2026 07:03
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #16. Stacked on the explicit-thinking control PR; this diff addresses deadlines only.
Raises the finite total_ms admission ceiling from 600,000 to 3,600,000 (one hour). Actual deadlines remain explicit in the workload and plan; existing example budgets are not silently raised. Idle must remain positive and no greater than total. Admission errors identify the offending deadline/buffer field and its bounds.
Collector clocks, cancellation/total/idle precedence, timeout statuses, retained partial evidence and no-retry policy are unchanged. In particular, raising total does not repair the reported 120-second idle timeout: silent prefill or queueing requires an explicitly appropriate idle budget. Longer deadlines can lengthen cooperative drain and are documented as a policy allowance, not a runtime guarantee.
Verified on this exact head: fmt, Clippy -D warnings, workspace tests and release build. Independent SIX and HELP approved the source/contract. CLI fixtures admit larger budgets without waiting an hour and check invalid relations and total versus idle expiry. No model-server calls or hour-long live qualification.