Add post-swap verification and automatic rollback to DeploymentHealthCheck subagent

[Copilot]

The DeploymentHealthCheck subagent swapped staging→production when response time exceeded baseline by 20%, but had no post-swap verification or rollback logic — leaving a degraded deployment in production when the swap made things worse.

Key changes to DeploymentHealthCheck.yaml

• Post-swap verification — After swap, waits 120s (App Service warm-up + App Insights ingestion latency), then re-queries App Insights with an explicit timestamp window anchored to SwapExecutedAt
• Automatic rollback — If post-swap avg response time is >10% worse than pre-swap, triggers an immediate rollback swap without approval; threshold is intentionally tighter than the 20% swap trigger to avoid false positive rollbacks
• Consistent KQL filters — All queries now use cloud_RoleName == 'sreproactive-vscode-39596' (explicit) instead of a negative not contains "staging" filter
• Evidence collection — ExecutePythonCode generates a timestamped CSV + bar chart PNG (Baseline / PreSwap / PostSwap) saved to /mnt/data/ and linked in the issue
• Enriched GitHub issue template — Matches the incident report format: baseline, pre/post swap windows with exact timestamps, deviation %, slot health check results, rollback status, evidence links, and the actual KQL queries used
• New tools added — GetCurrentUtcTime, WaitInMilliSeconds, ExecutePythonCode

# Auto-rollback logic (step 12)
- If PostSwapResponseTime > CurrentResponseTime * 1.10:
    execute: az webapp deployment slot swap ... (rollback)
    set: RollbackExecuted = true, PostSwapAssessment = "worse than pre-swap; rollback executed"
- Else:
    set: RollbackExecuted = false, PostSwapAssessment = "improved; swap successful"