Stay current without drama

Velocity ships regular minor releases and occasional breaking changes. Follow this checklist to keep clusters fresh while protecting availability.

Pre-flight

Sync these with your maintenance calendar so the upgrade never surprises on-call engineers.

Core flow

1. Fetch artefacts from the release page (checksums, signatures). Verify before distributing.
2. Run automation: unit tests, integration suites, and smoke tests for the edge runtime.
3. Deploy to staging using the same pipeline you’ll use for production.
4. Observe for at least one business cycle—watch handshake latency, downgrade percentage, and ticket rotation health.
5. Promote to production via canary, blue/green, or rolling upgrade depending on footprint.

Validate success

• velocity --version matches across all nodes.
• velocity-cli admin diagnostics --summary shows the expected schema version and ticket cadence.
• Config reload logs are clean—no deprecations or parse warnings.
• Dashboards show handshake success ≥ baseline and downgrades within thresholds.
• Synthetic probes (or Chaos tests) return within the agreed latency budget.

Rollback ready

1. Quiesce traffic with velocity-cli admin quiesce.
2. Redeploy the previous artefact or container image.
3. Restore the prior config bundle if schemas moved.
4. Resume traffic, monitor for 15 minutes, and file an incident ticket capturing root cause.

Keep a rollback runbook next to the upgrade plan so the team practises both paths.

Stay ahead of changes

• Subscribe to release emails or RSS.
• Watch issues tagged upgrade, breaking-change, and deprecation.
• Fold lessons into your automation—if a manual step keeps failing, script it and link the script from this guide.