v0.1.0

The first release of kubectl probes: a read-only kubectl plugin that shows, per workload and container, which startup, readiness, and liveness probes are configured, what their settings mean in practice, and what evidence of failure the cluster currently reports.

What it does

• Overview (kubectl probes, -A, -o wide): one row per container with the handler type and effective timing of each probe, ready pods, restarts, recent Unhealthy events, and a finding count. * marks a running pod whose spec has drifted from its pod template.
• Summary (--summary=show|only|hide): probe coverage and failure detection time across everything in scope, with the container at each percentile named.
• Inspection (kubectl probes deploy/api): one workload in full, with its probe configuration as a table, effective timing in sentences, per-pod runtime state, and failure evidence.
• Findings: seven opinionated rules, reported apart from the facts. docs/rules.md lists them, and --no-findings leaves them out.
• Manifests (-f): reads workloads that have not been applied yet.
• Machine-readable output (-o json, -o yaml).

Install

Download the archive for your platform below, unpack it, and put kubectl-probes on your PATH. checksums.txt has the SHA-256 of each archive. kubectl krew install probes will work once the plugin is listed in krew-index.

Worth knowing

• It only reads pods, their owners, and events. It never exercises a probe.
• The -o json and -o yaml schema is probes.kubectl.dev/v1alpha1 and may still change.
• FAILURES counts recent Unhealthy events only. The API server keeps those for about an hour by default.
• -l matches pod labels, not workload labels.