The Cluster Metrics page provides real-time visualization of your cluster’s resource usage. Connect a Prometheus data source to enable comprehensive metrics monitoring.
Overview
Ankra’s metrics visualization helps you understand your cluster’s health and resource consumption at a glance. View CPU usage, memory consumption, network throughput, disk I/O, and pod restart patterns—all from a single dashboard.Accessing Cluster Metrics
- Navigate to your cluster from the sidebar.
- Click Metrics in the cluster navigation.
- View real-time charts and summary cards.
Metrics require a configured Prometheus data source. See Prometheus Integration or Cluster Settings to configure your data source.
Summary Cards
At the top of the metrics page, summary cards show current resource utilization:| Card | Description |
|---|---|
| CPU Usage | Current CPU utilization vs total capacity |
| Memory Usage | Current memory consumption vs total capacity |
| Network Rate | Combined inbound and outbound network throughput |
| Pod Status | Running pods vs total pod count |
Metrics Charts
CPU Usage
Visualize CPU consumption across your cluster nodes over time. What it shows:- Per-node CPU usage lines
- Total cluster CPU trend
- Usage patterns over the selected time range
- Identify nodes with consistently high CPU usage
- Spot CPU spikes that correlate with deployments or traffic
- Plan capacity based on usage trends
Memory Usage
Track memory consumption across your cluster. What it shows:- Per-node memory usage
- Memory pressure indicators
- Historical memory trends
- Detect memory leaks in applications
- Identify nodes approaching memory limits
- Plan memory allocation for new workloads
Network I/O
Monitor network traffic flowing in and out of your cluster. What it shows:- Receive: Inbound network traffic per node
- Transmit: Outbound network traffic per node
- Throughput rates in bytes/second
- Identify network-intensive workloads
- Detect unusual traffic patterns
- Monitor data transfer costs
Disk I/O
Track disk read and write operations across nodes. What it shows:- Disk read throughput
- Disk write throughput
- I/O patterns per node
- Identify storage bottlenecks
- Monitor database disk activity
- Plan storage capacity
Pod Restarts
Monitor pod restart frequency to detect stability issues. What it shows:- Pod restart counts over time
- Restart patterns by namespace or workload
- Correlation with other events
- Detect crashlooping pods
- Identify unstable deployments
- Troubleshoot OOMKilled containers
Time Range Selection
Control the time window for metrics data:| Range | Use Case |
|---|---|
| Last 15 minutes | Real-time monitoring |
| Last hour | Recent activity review |
| Last 6 hours | Shift-based monitoring |
| Last 24 hours | Daily patterns |
| Last 7 days | Weekly trends |
| Custom | Specific time windows |
Changing Time Range
- Click the Time Range picker in the top-right corner
- Select a preset range or define a custom window
- Charts automatically update to show the selected period
Auto-Refresh
Keep metrics up-to-date with automatic refresh:| Interval | Description |
|---|---|
| Off | Manual refresh only |
| 10 seconds | Near real-time updates |
| 30 seconds | Balanced refresh rate |
| 1 minute | Low-overhead monitoring |
| 5 minutes | Background monitoring |
Manual Refresh
Click the Refresh button at any time to fetch the latest data immediately.Prometheus Configuration
Metrics require a connected Prometheus instance.Setting Up Prometheus
Install Prometheus
Deploy Prometheus to your cluster using a stack with the
kube-prometheus-stack add-on, or connect to an existing Prometheus instance.Default Prometheus URL
If usingkube-prometheus-stack deployed via Ankra:
Troubleshooting
”Prometheus Not Configured”
Cause: No Prometheus data source has been set up. Solution:- Go to cluster Settings → Metrics
- Configure your Prometheus URL
- Return to the Metrics page
”Unable to Load Metrics”
Cause: Connection to Prometheus failed. Solutions:- Verify Prometheus is running in your cluster
- Check the Prometheus URL is correct
- Ensure network connectivity between the Ankra agent and Prometheus
- Review Prometheus service account permissions
Missing Data for Some Metrics
Cause: Prometheus may not be scraping all required metrics. Solutions:- Verify
node-exporteris deployed for node metrics - Check
kube-state-metricsis running for Kubernetes metrics - Review Prometheus scrape configurations