Argo CD 3.6.0-rc1: AppSet Progressive Sync Metrics and CNPG Hibernation
Sources
Argo CD v3.6.0-rc1 introduces a few quality-of-life updates that target the "visibility gap" in large-scale AppSet deployments. Specifically, the addition of progressive sync metrics for AppSets is a necessary step for teams managing thousands of applications via a single template.
The Win: Progressive Sync Visibility
Until now, tracking the progress of a progressive sync (canary-style rollout) across an AppSet was largely a manual exercise in checking individual application statuses. By exposing workqueue rate limiter flags and specific metrics for app refreshes, Argo CD finally provides the telemetry needed to identify "stuck" rollouts before they become incidents. This reduces the operational toil of auditing large-scale changes.
The Catch: CNPG Hibernation Complexity
The introduction of hibernate and rehydrate actions for CloudNativePG (CNPG) clusters is interesting but introduces a new state machine into the GitOps flow. While "hibernating" a database to save costs in dev environments is a valid use case, it shifts the burden of state management from the database operator to the Argo CD action framework. If the rehydration trigger fails or the action timeouts, you're left with a "zombie" cluster that is neither fully off nor available for traffic.
Verdict
Upgrade for the AppSet metrics if you're running high-scale fleets. Exercise caution with the CNPG hibernation actions unless you have a robust alerting setup for your Argo CD action logs.