| name | observability-cli-guides |
| description | Curated map of the terminal query surface of every major observability backend. Use when observing a run - local or remote - to pick the stack's backend and learn how to authenticate and query its metrics, traces, logs, and profiles from a CLI - Grafana (gcx), Datadog (Pup CLI), Dynatrace (dtctl and DQL), Azure Monitor (az), AWS CloudWatch and X-Ray (aws), Splunk (splunk CLI and SPL). |
Observability CLI Guides
A selection map over the query CLIs of the major observability backends.
Pick the observed stack's backend, read its reference file, then follow the
linked official docs — the fetched page is the source of truth, not memory.
The method is the same everywhere: authenticate, discover what the
service emits, then query what you discovered.
Pick the backend
Each reference covers: setup and authentication, the discovery-then-query
commands per signal (metrics, traces, logs, profiles where the backend has
them), and Planning notes with the backend's coverage gaps and quirks.
For the local oddyssey stack (the Grafana case), the setup-local-stack
skill carries the ready-made gcx context — isolated config and datasource
UIDs. gcx is the stack's mandatory query CLI.
Rules
- Recommendations and commands must come from the reference's linked docs,
fetched — never from memory; CLIs move fast.
- Credentials come from the environment or the caller (env vars, secret
stores); never invent, echo, or store them.
- Planning notes are a snapshot (last verified 2026-08); the fetched
official page always overrides them.
- If the stack's backend is not in the table, say so and fall back to the
backend's documented REST API over
curl — the discover-then-query
method still applies.
