Troubleshooting

Start with the finding on the affected repository, runtime or task. Repair its cause before retrying, and keep the recorded result available for review.

Setup and readiness

Finding	Next action
GitHub access failed	Check App installation, repository access and connection status
Registry access failed	Repair push/pull access and rerun the connection tests
Registry cleanup warned	Inspect and retry cleanup independently of access tests
Policy or lock missing/incompatible	Commit supported Ayni configuration with the CLI
Certificate trust missing	Review and merge the linked trust-update PR
Existing environment mismatches	Inspect the finding, then explicitly choose Rebuild
Preparation exhausted its retries	Repair the cause, then use the explicit Retry or Rebuild action

Execution and delivery

Result	Next action
Quality failed	Read the findings and inspect the draft if one was delivered
Quality unavailable	Investigate missing check evidence or infrastructure; do not read it as passed
Deadline exceeded	Review consumed limits and cause before submitting another attempt
Publication failed	Retry publication when retained final work is available
Conflicting remote branch	Ask an operator to resolve it; recovery must not overwrite it
Cleanup pending	Allow owned-resource cleanup to finish; contact the deployment operator if it stalls

A deployment restart preserves admitted work and retained results. Avoid deleting retained data or credentials to make a retry appear successful. Contact your deployment operator for infrastructure failures, inaccessible administrator-managed integrations or recovery that needs stored data inspection.