Roll back at 19:40 without waking anyone
Pick the last good revision and redeploy it — the same button, whoever is on call.
“@here who has prod kubectl access? The 18:20 release is throwing 500s and I cannot roll it back myself”
What the wait costs
Thirty-five minutes of errors, an escalation, and one engineer’s evening — for an action that takes four seconds to perform.
In Fenwave
Deployments›checkout-api›History›Roll back
Capability
Deployment ManagerRolling back is not a difficult operation, it is a privileged one — which is why it ends up gated behind whoever holds a kubeconfig. Fenwave keeps every revision in the deployment history with the commit and author attached, and gives the roll-back button to the team that owns the service. ArgoCD performs the change, so the cluster state and the Git state stay in agreement.
How it works
- 01Open the service in Deployments and switch to History.
- 02Every revision lists its commit, author, duration and outcome.
- 03Select the last revision that was healthy and confirm the roll back.
- 04ArgoCD syncs the cluster to that revision; progress streams on screen.
- 05The roll back is itself recorded, so the timeline stays complete.