Skip to content

fix(snapshot): pause between Central snapshot deploy chunks - #313

Merged
youngledo merged 1 commit into
mainfrom
codex/snapshot-paced-deploy
Nov 13, 2026
Merged

youngledo merged 1 commit into
mainfrom
codex/snapshot-paced-deploy

Conversation

@youngledo

Copy link
Copy Markdown
Contributor

Summary

  • Run 43372565941 published about 30 modules, then Central Portal returned Cloudflare error 1015. The workflow retried after 90, 180, and 270 seconds. Each retry was rejected immediately and kept the block active. Sonatype's 429 guidance says retries during a block make the block worse.
  • Publish 20 modules at a time, with uploads serialized inside each module. Wait 10 minutes with no Central requests before the next chunk. If a chunk is still rate limited, retry only that chunk after the same quiet wait.
  • Use -nsu so a resumed chunk does not refresh snapshot metadata.

Validation

  • bash -n on the publish script
  • bash scripts/verify-supply-chain-workflows-test.sh
  • Reactor-id parsing checked against the failed job log

Impact

This publish takes longer: 136 modules are seven chunks with a 10-minute pause between chunks. Rerunning run 43372565941 still uses the old script. After this reaches main, start Snapshot manually. If today's earlier retries already lengthened the Central block, the first chunk waits through that block before it can succeed.

Continuous publishing trips Cloudflare error 1015 after about 30 modules,
and short retries refresh the block. Publish 20 modules at a time and wait
quietly before continuing or retrying the same chunk.
@youngledo
youngledo merged commit 9445b91 into main Nov 13, 2026
22 checks passed
@youngledo
youngledo deleted the codex/snapshot-paced-deploy branch November 13, 2026 08:58
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.

1 participant