Skip to content

Split actor retry exhaustion outcomes and increase backoff - #7414

Merged
apeacock1991 merged 2 commits into
mainfrom
sol/actor-retry-policy
Sep 18, 2026
Merged

apeacock1991 merged 2 commits into
mainfrom
sol/actor-retry-policy

Conversation

@apeacock1991

@apeacock1991 apeacock1991 commented Sep 17, 2026

Copy link
Copy Markdown
Contributor

Actor retry metrics currently combine attempt-limit and retry-budget exhaustion under one outcome. This splits them into attempts_exhausted and retry_budget_exhausted so operators can tell which constraint stopped a retry loop.

Raise the full-jitter base from 50ms to 500ms. There's evidence we're rapidly retrying and hitting the retry limit, so spacing them out more to see if that aids recovery rate.

@apeacock1991
apeacock1991 requested review from a team as code owners September 17, 2026 15:47
@ask-bonk

ask-bonk Bot commented Sep 17, 2026

Copy link
Copy Markdown
Contributor

@apeacock1991 Bonk workflow was cancelled.

View workflow run · To retry, trigger Bonk again.

Distinguish retry loops that reach the attempt limit from those that run out of retry-start budget. This lets observers identify which constraint ended the loop.
Raise the full-jitter base delay from 50ms to 500ms. The four retry caps now span 500ms through 4s, giving transient failures more time to clear without increasing the attempt limit or retry budget.
@apeacock1991
apeacock1991 merged commit b7add55 into main Sep 18, 2026
36 of 38 checks passed
@apeacock1991
apeacock1991 deleted the sol/actor-retry-policy branch September 18, 2026 10:01
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants