kube-saver and other Kubernetes tools¶
Choose by the data model and workflow you need. kube-saver offers local snapshots with modeled CPU/memory costs and review files. It does not ingest cloud invoices or maintain a historical usage model.
| Tool | Main role | Relationship to kube-saver |
|---|---|---|
| kube-saver | Local cost estimates, TUI, HTML snapshots, resource-change plans | Uses present metrics-server samples or estimated fallback |
| k9s | Terminal cluster navigation and operations | Useful alongside a cost-review tool |
| Goldilocks | Dashboard for VPA sizing recommendations | Uses VPA's recommendation model rather than kube-saver's snapshot heuristics |
| VPA | Resource recommendations and configurable updates | Can run recommendation-only or update modes; not always an automatic mutator |
| OpenCost | Ongoing Kubernetes cost allocation and cloud-cost monitoring | Open source and self-hostable; provides a broader allocation model |
OpenCost does not require a hosted SaaS account. A metrics stack or cluster components are different requirements from a hosted service. Consult each project's current installation docs before comparing deployment requirements.
Use kube-saver when¶
- You want a local review of requests versus present usage with configured rates.
- You need a self-contained HTML snapshot for internal review.
- You want local candidate patch commands to review and translate into GitOps.
Consider a different or complementary workflow when¶
- You need historical percentiles or ongoing automatic resource adjustments.
- You need cloud billing reconciliation, full infrastructure cost allocation, storage/network/GPU costs, or team chargeback.
- You need logs, exec, and cluster lifecycle operations from a terminal.
Reducing requests is not a measured cost saving. Evaluate representative load and whether the change can reduce billed capacity. See FAQ and safety.