Menu
Convert Prometheus alerting rules into Guance monitor JSON and validate all generated DQL.
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/GuanceCloud/ai-skills --skill alert-managerThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Showing SKILL.md
Based on SOC occupation classification
| name | alert_manager |
| author | liurui |
| description | Convert Prometheus alerting rules into Guance monitor JSON and validate all generated DQL. |
Use this skill when the user wants to convert Prometheus alerting rules into Guance monitor configuration.
The input must be an alerting rule definition with fields such as alert, expr, for, labels, and annotations. Do not treat Alertmanager routing configuration such as route, receivers, inhibit_rules, or mute_time_intervals as convertible monitor rules.
Before writing DQL, confirm the metric mapping:
| Required field | Purpose |
|---|---|
component | Output directory and stable monitor identity |
source_metric | Metric name in the source rule |
dataSource | Guance data source or measurement |
field | Guance field name |
groupBy | Monitor grouping tag |
fieldType | Field data type |
fieldFunc | Recommended aggregation function |
If any key mapping is missing, stop and ask for the missing mapping instead of inventing DQL.
monitor skill.checkers[].jsonScript.targets[].dql and checkers[].extend.querylist[].query.q for each rule.checkerOpt.rules and extend.rules synchronized.dqlcheck before final output.Map labels.severity to Guance statuses when possible: critical, warning, error, or info. If severity is missing, default to critical and state that the value was defaulted.
Stop instead of guessing when the expression uses multi-metric binary operations, complex and or or expressions, label_replace, histogram_quantile, complex subqueries, unknown recording rules, or rate semantics that cannot be safely mapped.