Skip to main content

avd-alert-kql-validation

Create and review KQL for the FederalAVD AVD Alerts add-on. Use when adding or changing scheduled query alerts, comparing alert logic with Microsoft AVD Insights workbooks, investigating query failures, or documenting intentional workbook deviations.

소스 정보

저장소
Azure/FederalAVD
최근 소스 활동
2026년 8월 18일 18:24
감지된 SKILL.md 언어
영어
스타
18
포크
19

설치 방법

기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.

소스 파일 검토

설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.

파일 탐색기
3 개 파일

SKILL.md 표시 중

SKILL.md
소스 지침 · 읽기 전용 미리보기
name
avd-alert-kql-validation
description
Create and review KQL for the FederalAVD AVD Alerts add-on. Use when adding or changing scheduled query alerts, comparing alert logic with Microsoft AVD Insights workbooks, investigating query failures, or documenting intentional workbook deviations.
argument-hint
[alert resource, KQL condition, or module]
# AVD Alert KQL Validation Use this workflow for log-based alerts under `deployments/add-ons/avdAlerts/`. ## Procedure 1. Read `deployments/add-ons/avdAlerts/README.md`, especially **Query Design and Validation**, and locate the corresponding operator response in `ALERT-RESPONSE.md`. 2. Identify the authoritative AVD Insights workbook for the condition. Use the official Microsoft workbook sources listed in [workbook sources](./references/workbook-sources.md). 3. Extract and structurally validate the repository queries: ```powershell & .github/skills/avd-alert-kql-validation/scripts/Test-AvdAlertQueries.ps1 ` -ModulePath deployments/add-ons/avdAlerts/modules ``` 4. To inspect standalone KQL, add `-OutputDirectory <temporary-folder>`. Do not commit generated extraction output. 5. Compare table selection, joins, event names or IDs, status filters, casts, deduplication, and aggregation with the workbook query covering the same condition. 6. Adapt interactive workbook queries for alert semantics: remove `render`, return rows suitable for the configured aggregation and threshold, and use a bounded time window. 7. Verify `overrideQueryTimeRange` is at least `windowSize` multiplied by `numberOfEvaluationPeriods` and that persistent-event behavior matches `autoMitigate`. 8. Preserve cloud compatibility. Do not introduce workspace-specific functions or preview-only KQL without confirming support in every target cloud. 9. Document every intentional deviation in the add-on README. The reason must describe alert semantics, noise suppression, cloud compatibility, or source-data behavior. 10. Build the affected Bicep entry template and update `ALERT-RESPONSE.md` when operator action, thresholds, or dimensions change.
GitHub에서 보기