redis/go-redis

Potential Redis Misconfiguration or Performance Issues Affecting ArgoCD Operations

Closed this issue · 1 comments

davidt-gh commented

Expected Behavior

Redis should operate seamlessly as a backend cache or data store for ArgoCD (or related services). It should consistently support high availability and handle operations without causing delays or warnings in the logs.

Current Behavior

The logs suggest potential misconfiguration or operational delays. Notable issues include:

  • Warnings about unassigned shards and possible delays before successful cluster assignment.
  • Repeated attempts to refresh app status and assign shards, which might indicate instability in the caching layer.

While the connection to Redis is not explicitly logged, its role in caching or storing app status might be impacted, leading to these behaviors.

Possible Solution

  1. Ensure Redis is configured correctly with appropriate resources and high availability.
  2. Validate Redis connection and timeout settings in ArgoCD or other components using Redis.
  3. Debug potential network or performance bottlenecks affecting Redis.

Steps to Reproduce

  1. Deploy ArgoCD with Redis as a backend component.
  2. Monitor logs during initialization and app status refresh.
  3. Note warnings or delays related to shard assignments or app state comparison.

Context (Environment)

argoproj/argo-cd#20824
application-controller.log

  • ArgoCD Version: v2.13.0+347f221
  • Redis Version: [Specify your Redis version]
  • Kubernetes Cluster: EKS (or specify)
  • Redis Deployment: [Clustered/Standalone]

Logs from ArgoCD suggest delays or issues during shard assignment and app state refresh, which might stem from Redis' behavior in caching or storing state data.

Detailed Description

The warning and informational logs in ArgoCD indicate delayed responses and misconfigurations during initialization and app state refresh. This could be linked to Redis' role in the stack. Investigating Redis configuration, connectivity, and resource allocation might help resolve the issue.

Possible Implementation

  • Add better monitoring and logging for Redis within the ArgoCD stack to isolate issues.
  • Validate Redis' health and ensure it is optimized for the current workload.
ndyakov commented

Based on the discussion in the argoproj/argo-cd#20824 it seems the issue was not related to the Redis library since it was fixed after updating Argo. Closing this.

  • 👍1