Skip to content

Retry apply-schema after a cold-start timeout - #43

Merged
ralyodio merged 1 commit into
profullstack:masterfrom
bartosz0718:retry-apply-schema-timeout
Sep 9, 2026
Merged

Retry apply-schema after a cold-start timeout#43
ralyodio merged 1 commit into
profullstack:masterfrom
bartosz0718:retry-apply-schema-timeout

Conversation

@bartosz0718

Copy link
Copy Markdown
Contributor

Summary

  • Retry a timed-out GET once in the shared client so a cold hosted board is not treated as dead (apply-schema from India missed 15s, then answered in ~1.4s). POST is not retried.
  • Keep the real HTTP status in stdio MCP. Create routes answer 201; reporting 200 made apply_to_job / post_job / post_update fail after a successful write.
  • Note the retry in README, /docs, llms.txt, OpenJob, and MCP tool text so agents that are not using this client still retry once.

Fixes #36

Test plan

  • node --test test/client.test.ts after pnpm build (GET retries once on timeout; POST does not; 201 is preserved)
  • agenticjobs schema <slug> against an idle hosted board still returns the schema
  • stdio MCP apply_to_job reports success after a 201 from POST /apply

…dio MCP.

An idle hosted board can miss a 15s GET, then answer in about a second. Agents that treat that as a dead endpoint skip applying. The CLI now retries GET once on timeout, and stdio MCP reports the real create status so apply_to_job no longer fails after a successful write.

Fixes profullstack#36
@ralyodio
ralyodio merged commit 226835d into profullstack:master Sep 9, 2026
4 checks passed
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.

apply-schema on hosted board can hang past 15s (then succeed)

2 participants