Progressive server summary as landing view per server

[erikdarlingdata]

Summary

When a user connects to a server, they face a tab bar and have to guess which tab matters. A server-level summary view — showing ranked problems with click-to-drill links — would eliminate the guessing and make the tool self-guiding.

Current Landing Experience

Dashboard

Top-level tabs: Overview, Queries, Plan Viewer, Resource Metrics, Memory, Locking, System Events. The Overview tab contains sub-tabs: Resource Overview, Daily Summary, Critical Issues, Default Trace, Current Configuration, Configuration Changes, Collection Health, Running Jobs. Resource Overview shows a snapshot but doesn't prioritize or rank problems.

Lite

Top-level tabs: Wait Stats, Queries, Plan Viewer, CPU, Memory, File I/O, TempDB, Blocking, Perfmon, Running Jobs, Configuration, Daily Summary, Collection Health. No overview or summary as the first tab — Wait Stats is the default landing tab.

Both apps require the user to already know where to look.

Proposed Design

A new server-level summary view that becomes the default landing experience when selecting/connecting to a server. It shows a ranked problem list with severity and click-to-navigate links:

Example Output

#1  BLOCKING STORM (Critical)
    14 events between 2:00-2:45pm, longest blocked 47 seconds
    Top blocker: SPID 128, dbo.usp_UpdateInventory
    [View Blocking Details →]

#2  CPU PRESSURE (Warning)  
    Avg 78% last hour, peak 96% at 2:12pm
    Top consumer: query 0x7A3F... (42% of CPU)
    [View Top Queries →]

#3  I/O LATENCY ELEVATED (Warning)
    TempDB write latency avg 28ms (threshold: 20ms)
    [View File I/O →]

#4  CONFIG CHANGE DETECTED (Info)
    MAXDOP changed from 4 to 8 at 1:45pm
    [View Configuration Changes →]

Click-to-Navigate Behavior

Each item links to the relevant tab with the time range pre-set to the incident window:

• Blocking → Blocking tab, Blocked Process Reports sub-tab, slicer set to incident window
• CPU → Queries tab, Top Queries sub-tab, filtered to incident window
• I/O → File I/O tab (Dashboard) or File I/O tab (Lite)
• Config → Configuration Changes (Dashboard Overview sub-tab) or Configuration tab (Lite)

This pairs directly with issue #684 (Critical Issues investigate action) — same navigation pattern, surfaced proactively.

Relationship to Existing Views

Dashboard

• Resource Overview already shows a snapshot but doesn't rank or prioritize
• Critical Issues shows detected problems but as a flat grid with no drill-down
• Daily Summary shows aggregated health but at day granularity
• The summary view would synthesize these into a single ranked, actionable entry point

Lite

• No equivalent exists today
• The analyze_server engine in the MCP layer already produces ranked findings with severity — this is its visual equivalent
• The ErikAI analysis engine (in progress) could feed this view directly

Design Notes

• This doesn't replace existing tabs — it becomes the front door
• Tabs become drill-down destinations reached from the summary, not guessing targets
• Should auto-refresh with the same cadence as other views
• Empty state (no issues detected): show a green "all clear" with key metrics summary
• Data sources: combine critical issues, wait stats, CPU, blocking counts, config changes, I/O stats — all already collected
• Both apps should implement this, though the data sources differ (Dashboard reads from SQL Server, Lite reads from DuckDB)

[erikdarlingdata]

Implementation order: Lite first, then port to Dashboard once the pattern is validated. This avoids the complexity of iterating on both apps simultaneously.

[erikdarlingdata]

Implementation Notes

This should be implemented together with #691 (USE/RED/Golden Signals framing). The progressive summary is the UI surface for the framework vocabulary. See the design sketch on #691.

Key decisions:

• Combined implementation: Frame server analysis output using USE/RED/Golden Signals vocabulary #691 provides the framework structure, Progressive server summary as landing view per server #689 provides the UI
• Replaces the default landing tab (Wait Stats in Lite, Resource Overview in Dashboard)
• All data sources already exist — this is aggregation + presentation, not new collection
• Navigate actions reuse the pattern from Add Investigate action to Critical Issues that navigates to relevant view #684 (already built)
• Lite first, then Dashboard port

Dependencies satisfied:

• Add Investigate action to Critical Issues that navigates to relevant view #684 (Critical Issues investigate) — merged, provides the navigation pattern
• Add time range slicers to Blocking, Deadlocks, and Default Trace Events views #681 (slicers) — merged, drill-down targets have slicers
• Add chart point drill-down to CPU, Memory, TempDB, and Blocking charts #682 (chart drill-down) — merged, chart-level investigation available from target tabs

[erikdarlingdata]

Implemented in PR #719. Lite now has an Overview tab as the default landing view with a 2x2 resource chart grid (CPU, Memory, File I/O, Wait Stats). Dashboard already had this via Resource Overview. The USE/RED/Golden Signals framing (#691) remains open as a future enhancement to organize and label these signals.