From gcx
Checks SLO health, budgets, burn rates, and trends using gcx CLI commands. Displays overviews of all SLOs, specific status, timelines, and graphs for Grafana Cloud.
npx claudepluginhub grafana/gcx --plugin gcxThis skill is limited to using the following tools:
Check SLO health, budget consumption, and trends. Route to investigation or optimization as needed.
Analyzes SLO performance trends over 28 days using gcx, computes SLI statistics like mean and std dev, and recommends data-backed improvements for objectives, alerting, labels, or windows.
Provides Prometheus queries and templates for SLO/SLI definitions on availability/latency, error budget calculations, and burn rate alerting for service reliability.
Guides defining SLOs, selecting SLIs, and implementing error budget policies for service reliability, alerting, and balancing velocity.
Share bugs, ideas, or general feedback.
Check SLO health, budget consumption, and trends. Route to investigation or optimization as needed.
gcx commands — do not call Grafana APIs directly-o json for agent processing, default format for user displayIf the user asks about all SLOs (no specific UUID):
gcx slo definitions list
This shows UUID, name, target objective, window, and status for all SLOs.
Then get the health summary:
gcx slo definitions status
This shows SLI, error budget, and health status for all SLOs in a table.
If the user asks about a specific SLO (UUID or name provided):
gcx slo definitions status <UUID> -o wide
The -o wide output includes additional columns: BURN_RATE, SLI_1H, and SLI_1D, which give a richer picture of recent performance.
| Status | Meaning |
|---|---|
| OK | SLI >= objective. SLO is healthy. |
| BREACHING | SLI < objective. Error budget is being consumed. |
| NODATA | No Prometheus metrics from recording rules. |
| Creating | SLO provisioning in progress. |
| Updating | SLO update in progress. |
| Deleting | SLO deletion in progress. |
| Error | SLO in error state — check the SLO configuration. |
NODATA handling: Recording rule metrics may not be available if the destination datasource is misconfigured or recording rules are not evaluating. Note this to the user and suggest checking the destination datasource configuration.
Show the timeline when:
All SLOs timeline:
gcx slo definitions timeline --from now-7d --to now
Specific SLO timeline:
gcx slo definitions timeline <UUID> --from now-7d --to now
Use --since as a shorthand when a single duration is more natural:
gcx slo definitions timeline <UUID> --since 7d
The timeline command renders a graph by default — this is the preferred output for users.
Adjust the time range based on the SLO window:
--from now-7d --to now--from now-28d --to nowWhen the user asks about SLO reports or wants combined SLO health from the reports subsystem:
gcx slo reports status
Or for a specific SLO:
gcx slo reports status <UUID>
BREACHING SLOs: After showing the timeline, suggest slo-investigate:
SLO
<name>is BREACHING (SLI: , objective: ). For a deep-dive investigation — raw metrics, dimensional breakdown, alert rules — use the slo-investigate skill.
User asks about improvements: Route to slo-optimize:
For trend analysis, objective tuning recommendations, and alerting sensitivity review, use the slo-optimize skill.
All-SLOs overview:
SLOs: <total> total, <n> OK, <n> BREACHING, <n> NODATA
[Table from gcx slo definitions status]
[If any BREACHING: show timeline graph]
[If BREACHING: suggest slo-investigate for each BREACHING SLO]
Specific SLO status:
SLO: <name>
Status: <OK|BREACHING|NODATA>
SLI: <value> (objective: <value>, window: <value>)
Error budget remaining: <value>
Burn rate: <value>
SLI_1H: <value> | SLI_1D: <value>
[If trend requested or BREACHING: show timeline graph]
[Routing suggestions as applicable]
Use minimal formatting. Lead with the status and key metrics. No excessive bold text.
Collect all errors and report at the end of the workflow — do not interrupt the workflow for non-fatal errors.
| Error | Action |
|---|---|
gcx slo definitions list returns empty | Report "No SLOs found in this context." Check context with gcx config view. |
gcx slo definitions status returns NODATA for all SLOs | Note that recording rule metrics are unavailable. Suggest checking the destination datasource configured on each SLO definition. |
gcx slo definitions status <UUID> — UUID not found | List all SLOs to help the user identify the correct UUID. |
gcx slo definitions timeline fails | Note the failure and continue. Timeline is supplementary. |
gcx slo reports status fails | Note the failure. Reports status is optional context. |
| Auth errors | Check context configuration: gcx config view. Ensure the server URL and credentials are set. |