| name | k-logs |
| description | Tail logs for a k3s-hosted service (traefik, grafana, prometheus, freshrss, node-exporter). Thin wrapper around kubectl logs that handles namespace + label lookup so you don't have to remember them on mobile. |
k-logs
Tail logs from a k3s service. Saves typing kubectl logs -n <ns> -l app=<x> --tail=N on mobile.
Required input
- service (string): one of the third-party k3s services. Common:
traefik, grafana, prometheus, freshrss, node-exporter.
- lines (int, optional): how many lines to tail. Default 100. Cap at 500 for mobile readability.
Steps
- Find the namespace + label. Run
k get pods -A -l app=<service> to locate the pod. If the label doesn't match (some charts use app.kubernetes.io/name), try k get pods -A | grep <service> and infer.
- Tail.
k logs -n <ns> -l <label>=<service> --tail=<lines>. If multiple pods match (e.g. DaemonSet), the output is interleaved — that's expected.
- Report. Output the tail directly. If it's >500 lines or very noisy, suggest filters (e.g.
| grep ERROR) rather than dumping.
When to not use this
- Host-Docker apps (blog, hub, stickies, delta_neutral, tbitt) — those are in Docker Compose, not k3s. Use
docker logs <container> instead.
- When you need to follow logs indefinitely. This skill is for a one-shot tail. For
-f follow mode, just call k logs ... -f directly; don't invoke this skill.