- Set replicas: 2 with maxUnavailable: 0 / maxSurge: 1 on both
deployments so rolling updates never drop below 2 ready pods
- Add PodDisruptionBudgets (minAvailable: 1) to protect against
node drains and cluster maintenance
- Add Fly.io cache purge step to cv-deploy workflow (docs already
had this) so CV deploys don't serve stale cached content
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>