Production Argo CD with the HA manifests
What the HA installation changes compared with the default manifests: replicas per component, three-node anti-affinity, Redis HA with Sentinels, API server replica settings, and the IPv6-only limitation.
Argo CD high availability and scaling
Keeping Argo CD responsive as applications, repositories and clusters grow: HA manifests, controller sharding, repo-server and Redis tuning, monorepo performance and disaster recovery.
Argo CD application controller sharding
Spreading clusters across application controller replicas: legacy, round-robin and consistent-hashing sharding, pinning a cluster to a shard, dynamic distribution, and signs that one controller is overloaded.
Tuning the Argo CD repo-server for large Helm and Kustomize workloads
Fixing slow or OOM-killed manifest generation: parallelism limits, tool exec timeouts, Git retry counts, manifest cache lifetime, directory manifest size limits, and scaling repo-server replicas.
Argo CD performance with monorepos
Stopping every commit to a large repository from regenerating every application: the manifest-generate-paths annotation, fully qualified target revisions, shallow clone trade-offs, webhooks and refresh jitter.
Running Argo CD with a thousand or more applications
Settings that matter at scale: status and operation processors, repo-server timeouts behind Context deadline exceeded errors, application tree sharding in Redis, reconciliation timeouts and when to split into shards or instances.
Reducing Git polling load with Argo CD webhooks
Replacing three-minute polling with push notifications from Git: webhook setup per provider, shared secrets, payload limits, webhook refresh jitter, and the branch and tag naming gotcha.
Argo CD backup and disaster recovery
Recovering Argo CD after losing its cluster: what argocd admin export captures, matching image versions, namespace pitfalls, what Git already restores, and rehearsing a restore with RTO targets.
Scaling the Argo CD API server and Redis
Sizing the API server for many UI and CLI users and Redis for many applications: replica counts and the login rate limit, Redis HA layout, and controller-to-Redis traffic.
