Fix 6 verified Lite bugs from code review

[erikdarlingdata]

Summary

• Anomaly detector rate normalization: Wait/blocking/deadlock spike detection now compares per-hour rates instead of raw totals across different-length windows (24h baseline vs 4h analysis). Previously a "5x spike" threshold required ~30x the actual hourly rate to trigger — anomaly detection was effectively non-functional.
• Poison wait UTC fix: Replaced the only NOW() call in the codebase with parameterized DateTime.UtcNow to match UTC-stored collection_time
• Daily summary aggregation: top_wait_type now uses GROUP BY wait_type ORDER BY SUM() instead of picking the single highest individual delta sample
• FinOps High Impact filters: Routes data through DataGridFilterManager instead of setting ItemsSource directly — column filters were broken on this grid
• Plan viewer accuracy ratio: Divides ActualRows by ActualExecutions before comparing to EstimateRows — operators with many executions no longer show false "1000x off" warnings
• Settings Test Email: Saves/restores App SMTP values around the test instead of mutating them permanently. Cancel no longer leaves test values in effect or persists password to credential store.

Test plan

• Run Lite against a server with active workload, verify anomaly detection triggers on real spikes
• Verify daily summary shows the correct top wait type (highest aggregate, not highest single sample)
• Open FinOps > High Impact Queries, apply a column filter, verify it works
• Open a plan with Nested Loop joins, verify accuracy % in node visuals matches edge tooltips
• Open Settings > SMTP, click Test Email, then Cancel — verify SMTP settings revert

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved anomaly detection accuracy for wait times, blocking, and deadlock analyses by normalizing metrics across different reporting windows.
  • Corrected execution plan visualization to display actual rows per execution versus estimated rows.
  • Enhanced identification of top wait types in daily performance summaries.
  • Fixed SMTP settings to preserve original values correctly during test email operations.

[coderabbitai]

Caution

Review failed

Pull request was closed or merged during review

📝 Walkthrough

Walkthrough

Updates anomaly detection to normalize metrics by window lengths, adjusts UI data flows through filter managers, refines row-per-execution visualization calculations, modifies SQL queries for top-wait-type and timestamped filtering, and introduces state preservation for SMTP settings during email testing.

Changes

Cohort / File(s)	Summary
Anomaly Detection Normalization Lite/Analysis/AnomalyDetector.cs	Reworked wait, blocking, and deadlock anomaly comparisons to normalize metrics by window lengths (per-hour rates). Introduces baselineHours/currentHours calculations, derives per-hour ratios (baselineRate/currentRate), and replaces absolute ratio logic with normalized rate-based ratios. Threshold checks updated accordingly across all anomaly types.
UI Control Data Flows Lite/Controls/FinOpsTab.xaml.cs, Lite/Controls/PlanViewerControl.xaml.cs	FinOpsTab replaces direct grid data-source assignment with filter manager update call. PlanViewerControl computes actual rows per execution (ActualRows/ActualExecutions) and uses this for accuracy ratio calculation and display instead of raw ActualRows.
Data Service Query Updates Lite/Services/LocalDataService.DailySummary.cs, Lite/Services/LocalDataService.WaitStats.cs	DailySummary changes top\_wait\_type query to group by wait\_type and order by aggregated SUM(delta\_wait\_time\_ms). WaitStats converts relative timestamp filter (NOW() - INTERVAL) to parameterized DateTime.UtcNow-based filtering for deterministic window handling.
Settings State Preservation Lite/Windows/SettingsWindow.xaml.cs	Introduces temporary backup (origServer/origPort/origSsl/origUsername/origFrom/origRecipients) of current SMTP settings before email test, then restores them in a finally block to ensure only explicit Save operations persist changes.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• Add anomaly detection — baseline comparison for acute deviations (#589) #606 — Directly updates the same anomaly detection comparison logic in AnomalyDetector.cs with per-hour rate normalization and ratio threshold adjustments for wait, blocking, and deadlock analyses.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 57.14% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'Fix 6 verified Lite bugs from code review' accurately summarizes the main objective of the changeset, which addresses six specific bug fixes across the Lite codebase identified from code review.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/lite-code-review-findings

📝 Coding Plan

• Generate coding plan for human review comments

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}