Use this dashboard to monitor your Flux control plane across reconciliation throughput and errors, reconcile duration, controller workqueues, runtime resource usage, and Kubernetes API client traffic.
Dashboards → + New dashboard → Import JSON
What This Dashboard Monitors
- Overview: Reconciliations per second, reconcile errors, active workers, workqueue depth, reconcile duration P99, and Kubernetes API request rate
- Reconciliation: Reconciliation rate by result, by controller, and by Flux kind, plus error rate by controller
- Reconciliation Duration: Average duration by kind, P50/P95/P99 percentiles, P99 by kind and by resource, and controller reconcile time P99
- Workqueue: Workqueue depth, add rate, retry rate, queue duration P99, and unfinished work by controller
- Resources: CPU usage, working-set memory, and goroutines by controller
- Kubernetes API (rest_client): Request rate by status code, method, and host, plus the non-2xx error rate
Metrics Included
Overview
- Reconciliations / sec: Total reconciliation throughput across controllers
- Reconcile Errors (range): Reconciliation errors across the selected range
- Active Workers: Active reconcile workers per controller
- Workqueue Depth: Pending items in controller workqueues
- Reconcile Duration p99: 99th percentile reconciliation duration
- K8s API Reqs / sec: Kubernetes API client request rate
Reconciliation
- Reconciliation Rate by Result: Reconcile rate split by success and error
- Reconciliation Rate by Controller: Per-controller reconcile rate
- Reconcile Error Rate by Controller: Per-controller reconcile error rate
- Reconciliations by Flux Kind: Reconcile rate grouped by Flux resource kind
Reconciliation Duration
- Avg Reconcile Duration by Kind: Mean reconcile duration per Flux kind
- Reconcile Duration p50 / p95 / p99: Reconcile duration percentiles
- Reconcile Duration p99 by Kind: 99th percentile reconcile duration per Flux kind
- Reconcile Duration p99 by Resource: 99th percentile reconcile duration per resource
- Controller Reconcile Time p99 by Controller: Controller-runtime 99th percentile reconcile time per controller
Workqueue
- Workqueue Depth by Controller: Pending workqueue items per controller
- Workqueue Add Rate by Controller: Rate of items added to each controller's workqueue
- Workqueue Retry Rate by Controller: Rate of workqueue retries per controller
- Workqueue Queue Duration p99: 99th percentile time items wait in the queue
- Workqueue Unfinished Work Seconds: Seconds of in-progress work not yet finished
Resources
- CPU Usage by Controller: Per-controller CPU usage
- Memory (Working Set) by Controller: Per-controller working-set memory
- Goroutines by Controller: Goroutines per controller
Kubernetes API (rest_client)
- K8s API Request Rate by Status Code: Kubernetes API request rate grouped by HTTP status code
- K8s API Request Rate by Method: Kubernetes API request rate grouped by HTTP method
- K8s API Request Rate by Host: Kubernetes API request rate grouped by host
- K8s API Error Rate (non-2xx): Rate of non-2xx Kubernetes API responses
Dashboard Variables
- k8s_cluster_name: Filter by Kubernetes cluster
- kind: Filter by Flux resource kind
- k8s_namespace: Filter by Kubernetes namespace
