Whoops! How Not to Accidentally Delete Everything.
summary
Explains exactly how Argo CD deletion works: finalizers trigger foreground cascade deletes, CRD deletion removes all applications, and `--force` can delete-and-recreate causing mass wipes; an ApplicationSet generator outage or reverted feature in late March caused thousands of apps to render empty and be pruned. Shows concrete mitigations: set preserveResources on deletion or orphan propagation policy, disable automated sync or use sync windows, migrate by creating a second Argo instance and non‑cascading delete.
AI-written summary · about
chapters (14)
- 0:00Avoiding accidental deletions
- 2:33Apps deleted unexpectedly
- 3:15Argo CD deletion mechanics
- 8:08Owner references vs finalizers
- 10:13Renaming apps in Argo CD fails
- 12:41Safe renaming of app sets
- 13:03CRD deletion dangers during upgrades
- 14:52Safe Argo CD upgrades
- 15:57Handling finalizers to prevent loss
- 17:03Orphan policy for protection
- 18:07App sets error resilience pitfalls
- 20:21Preserve resources on deletion
- 21:45Sync windows and prune ops
- 23:09Summary
related talks
37mEffective Disaster Recovery: The Day We Deleted Production
28mArgo CD Maintainer's Panel
05mProject Lightning Talk: Argo CD at Scale: 5 Features You Should Not Miss!
25mRUD: Rapid Unscheduled Deletion of Critical Infrastructure Workloads
05mWe accidentally deleted production, but we made it not matter all that much
48mHow a Couple of Characters (and GitOps) Brought Down Our Site
18mUnveil the Secret Ingredients for Argo CD in the Enterprise-Scale
36mMulti-cluster, multi-apps, multi-value deployments using Argo CD Application Sets