microsoft / microsoft/GitHub-Copilot-for-Azure
azure-deploy: Integrate post-deployment log fetching with existing diagnostics tools
- Dominant language
- Python
- Stars
- 250
- Forks
- 204
- Avg merge
- 1d 12h
- Merged PRs (30d)
- 67
Description
## Problem
After `azd up`, if the deployed app has a bug, the agent has no way to fetch logs and inform the user. In one test case, the agent hit an app bug post-deployment and timed out trying to fix it instead of reporting the error.
## What exists
- **`azure-diagnostics`** can query logs via `mcp_azure_mcp_monitor` (KQL) and `mcp_azure_mcp_applens` (root cause analysis) — but needs app name/resource group explicitly.
- **`azure-deploy`** has verify recipes with `az containerapp logs show` and `az webapp log tail` — but only as a manual fallback when health checks fail.
## What's missing
- No handoff from deploy → diagnostics. After `azd up`, deployed resource info isn't passed to monitoring tools.
- No combined flow: deploy → verify → fetch logs.
- The two skills are completely independent — no cross-references, no shared context.
## Expected behavior
If post-deployment health check fails, the agent should use existing diagnostics tools to **fetch and surface app logs** to the user — not attempt to fix the bug.
Contributor guide
Assessment
This issue has not been assessed yet.