# Diagnose Azure Production Issues

Azure production incidents can span services, logs, metrics, credentials, and network paths. This skill guides structured triage with Azure Monitor, AppLens, KQL, CLI, and service-specific checks.

## Install

```bash
npx skillstore add microsoft/azure-diagnostics
```

## Metadata

- Status: approved
- Slug: microsoft-azure-diagnostics
- Version: 0.0.0-placeholder
- Author version: 0.0.0-placeholder
- Skillstore revision: r2
- Version status: valid
- Tree hash: 0df340b78b508bf68ade37a023f20a105b908ab124957f0dee9b26bbbc524627
- Author: microsoft
- GitHub username: microsoft
- License: MIT
- Repository: https://github.com/microsoft/github-copilot-for-azure/tree/main/plugin/skills/azure-diagnostics/
- Ref: ebdfe608f5de2b66ff37ab4af12af8ac4f5e8006
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: network, filesystem, env\_access, external\_commands
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/microsoft-azure-diagnostics
- Manifest: https://skillstore.pages.dev/api/skills/microsoft-azure-diagnostics/manifest

## Capabilities

- Routes incidents to playbooks for App Service, Container Apps, Functions, AKS, Event Hubs, and Service Bus.
- Guides resource health, activity log, metric, application log, and KQL investigations.
- Provides Azure CLI and MCP templates for gathering evidence and applying selected remediations.
- Maps common symptoms to likely causes and practical remediation options.
- Covers identity, connectivity, retries, checkpointing, health probes, deployments, nodes, pods, ingress, and DNS.

## Use Cases

- Triage a production incident: Correlate resource health, recent changes, logs, and metrics before selecting a remediation.
- Debug application failures: Investigate deployment errors, health probes, cold starts, runtime settings, and dependency failures.
- Diagnose platform connectivity: Analyze AKS, Event Hubs, and Service Bus identity, networking, scaling, and client behavior.

## Prompt Templates

### Start basic incident triage

```
Diagnose this Azure issue: [symptom]. The affected service is [service], the resource is [resource], and the problem began at [time].
```

### Build an evidence plan

```
Build a diagnostic plan for [resource]. Use [available logs or metrics], separate evidence from hypotheses, and list only read-only commands first.
```

### Analyze a service failure

```
Troubleshoot [failure] in [Azure service]. Compare identity, configuration, networking, resource health, recent changes, and service-specific failure patterns.
```

### Correlate a complex incident

```
Investigate [incident] across [resources]. Correlate deployments, resource health, KQL results, identity, dependencies, and network paths. Rank root causes by evidence.
```

## Limitations

- Requires valid Azure authentication, correct RBAC permissions, and accurate resource identifiers.
- Does not replace Azure support or confirm platform outages beyond available health data.
- Some examples can change resources and require review before execution.
- Coverage is limited to the listed Azure services, SDKs, and documented scenarios.

## Best Practices

- Start with resource health and recent changes before collecting detailed logs.
- Use read-only queries first, then request approval before any resource change.
- Redact secrets, connection strings, tokens, customer data, and sensitive log fields from shared outputs.

## Anti Patterns

- Do not run combined remediation commands before replacing placeholders and validating scope.
- Do not treat a single error message as root cause without supporting logs or metrics.
- Do not expose credentials or paste unrestricted production logs into prompts.

## Security Audit

- Audited at: 2026-07-23T18:34:19.639\+00:00
- Summary: All 58 static alerts are contextual false positives from Markdown, placeholder endpoints, environment-selection examples, and diagnostic commands. However, SKILL.md contains a high-severity authority claim that resembles prompt injection and could improperly influence an agent's instruction hierarchy.

## Stats

- Views: 120
- Downloads: 12
- Favorites: 0
- Popularity score: 0