Optimizer
The Kubernetes Optimizer analyzes your container resource requests and limits against actual usage, then recommends right-sizing changes. Use it to reduce cloud costs by reclaiming over-allocated memory and CPU.
Getting Started
Navigate to Kubernetes → Optimizer in the sidebar. The page displays a table of containers with optimization recommendations.
Optimizer Table
The table toggles between Memory and CPU tabs:
| Column | Description |
|---|---|
| Container | Container name |
| Cluster | Cluster the container runs in |
| Namespace | Kubernetes namespace |
| Pods | Number of pods running this container |
| Memory Peak (30d) | Peak memory usage over the last 30 days |
| Memory Request / Limit | Current configured request and limit |
Summary Banner
At the top, a summary shows the total reclaimable resources:
- Total Reclaimable — Aggregate memory or CPU that can be saved.
- Memory — Total reclaimable memory in GB.
- CPU — Total reclaimable CPU in cores.
Filters
Use the left-hand panel to narrow the view:
- Cluster — Filter by cluster.
- Namespace — Filter by namespace.
- Container — Search for a specific container.
How It Works
- Oodle continuously collects resource usage metrics from your clusters.
- The Optimizer compares the peak usage over a 30-day window against the configured requests and limits.
- Containers where requests significantly exceed actual peak usage are flagged as over-provisioned.
- Recommendations are generated to right-size the requests and limits.
Use Cases
- Cost optimization — Identify containers requesting far more memory or CPU than they actually use.
- Capacity planning — Understand real resource consumption patterns to plan future scaling.
- Right-sizing — Set requests and limits based on observed peak usage rather than guesswork.
Best Practices
- Review the 30-day peak before downsizing — Short-lived spikes may not appear in shorter windows.
- Apply changes gradually — Adjust one namespace or workload at a time and monitor for issues.
- Leave headroom — Do not set limits exactly at peak; add a safety margin (e.g. 20%) to handle load variations.
Support
If you have any questions or need assistance, please contact us via our help chat app available on the Support link in the sidebar, or by reaching out to support@oodle.ai.