Skip to content

ci(actions): harden hosted-runner resilience - #149

Merged
jsugg merged 3 commits into
masterfrom
fix/ci-resilience
Jul 29, 2026
Merged

ci(actions): harden hosted-runner resilience#149
jsugg merged 3 commits into
masterfrom
fix/ci-resilience

Conversation

@jsugg

@jsugg jsugg commented Jul 29, 2026

Copy link
Copy Markdown
Owner

Summary

  • harden local Edge startup and bound idempotent Java navigation
  • move CI Docker startup into a tested retry/readiness helper
  • reduce nightly registry bursts with jitter and matrix concurrency limits
  • upgrade Node/Python setup and artifact actions to supported runtimes
  • align Dependabot commit scopes and CI-guide policy

Root cause addressed

Nightly run 30190689132 failed before checkout when one hosted runner exhausted
Docker Hub pull retries for the pinned application image. Job-level service
containers start before workflow steps, so the repository could neither tune
retries nor emit useful diagnostics.

Resilience design

  • four bounded image-pull attempts with exponential backoff and jitter
  • digest validation and --pull=never container startup
  • explicit website and Selenium Grid readiness checks
  • grouped diagnostics and cleanup on startup failure
  • no retries around assertions or state-changing test actions
  • at most three concurrent legs per nightly matrix

Hosted validation

PR checks and a branch-scoped nightly dispatch will be required before merge.

jsugg added 3 commits July 28, 2026 22:55
Bound page loads and retry only idempotent navigation timeouts.
Transient hosted-runner browser stalls now fail quickly and diagnostically.
Move Docker startup into a tested workflow step. Registry pulls gain
bounded retries, jitter, readiness checks, and actionable diagnostics.
@jsugg

jsugg commented Jul 29, 2026

Copy link
Copy Markdown
Owner Author

Hosted validation is complete: the full 20-job nightly matrix passed on this branch, including Java local Edge and Java Grid Firefox—the two previously observed failure paths. Run: https://github.com/jsugg/the-internet-tests/actions/runs/30415496399

@jsugg
jsugg merged commit 9306e4b into master Jul 29, 2026
37 checks passed
@jsugg
jsugg deleted the fix/ci-resilience branch July 29, 2026 02:06
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