Reports¶
k13d currently generates cluster reports from the Web UI.
What Reports Include¶
Reports can include these sections:
- Nodes: node readiness, cordon state, pressure warnings, taints, capacity and allocatable values
- Namespaces: namespace activity and workload counts
- Workloads: pods, deployments, services, and top container images
- Events: recent warning events
- Security: built-in pod / RBAC / network / privilege signals
- Security Full: extended scan when the security scanner is available
- FinOps: heuristic compute-cost analysis and rightsizing guidance
- Metrics: historical cluster metrics when the collector is enabled
- AI Analysis: optional narrative summary from the configured LLM
Generate A Report¶
- Open Reports in the Web UI.
- Select the sections you want.
- Optionally enable AI Analysis.
- Preview in-browser or download the report.
The selected sections now control the exported HTML/CSV output as well. If you do not select a section, it is omitted from the generated report.
Output Formats¶
k13d currently supports:
- HTML: best for human-readable reports and browser preview
- CSV: tabular export for spreadsheets and follow-up analysis
- JSON: raw structured data
There is no standalone k13d report CLI command and no built-in PDF or Markdown export in the current binary. For PDF, download HTML and use your browser's Print → Save as PDF flow.
FinOps Notes¶
The FinOps section is intentionally a heuristic estimate, not a cloud invoice.
- It focuses on compute-style cost signals from running pod requests.
- If live pod metrics are available, k13d uses them to improve usage and efficiency fields.
- If metrics-server is unavailable, k13d falls back to request-derived estimates and labels the result accordingly.
- Direct provider charges such as control-plane fees, storage classes, egress, committed-use discounts, and reserved capacity are not modeled precisely.
Use the FinOps section as a prioritization tool:
- find namespaces driving the largest share of estimated spend
- identify pods missing requests/limits
- spot underutilized workloads when live metrics exist
- review LoadBalancer sprawl for direct savings opportunities
Node Health Checks¶
The node section is meant to be operationally useful, not just inventory.
Each node report includes:
- Ready / NotReady state
- Cordoned (
Unschedulable) state - pressure conditions such as
MemoryPressure,DiskPressure, andPIDPressure - network availability warnings
- taints
- capacity and allocatable CPU / memory values
This makes the report usable as both a lightweight cluster assessment and a handoff artifact when a node issue is suspected.