[Feature Request] Display Aggregate Total Trace Duration in Search Results Table (Lucene Mode) #3038

Description

Currently, when searching for a specific traceId using Lucene Mode in the Logs/Traces search interface, the results table outputs a flat list of all matching individual spans. While the table shows the standalone duration for each span (sub-operation), there is no clear indicator of the absolute end-to-end total execution time (wall-clock duration) for the entire trace on the screen.

Because spans frequently run in parallel (concurrently) or are nested hierarchically as parent-child processes, users cannot simply sum the individual row values together without aggressively double-counting execution time. To find the true total duration, a user is forced to switch to the visual Waterfall Trace View or manually subtract the earliest span start timestamp from the latest span end timestamp.

Proposed Solution

Introduce a visible "Total Trace Duration" summary metric or structural row block directly within the search table view when filtering by a single traceId.

Suggestible Option: Visually pin or highlight the calculated "Root Span" at the top or bottom of the results table list so its comprehensive runtime is instantly distinct from its internal child spans.

Why This Matters (Business & Engineering Importance)

1. Drastically Reduces Time-to-Resolution (TTR)

When a production outage or latency spike occurs, every second counts. On-call engineers frequently pull up a specific traceId from an error log to quickly assess how slow the transaction was. Forcing engineers to switch screens or perform manual timezone/millisecond math to calculate total wall-clock latency slows down triage under pressure.

2. Avoids Misleading Metrics & Parallel Overhead Confusion

A flat duration column can easily lead to misinterpretation. For instance, seeing multiple parallel rows showing 2123ms and 2207ms can lead an engineer to assume a bottleneck took ~4.3 seconds, when the reality is they ran concurrently and finished in 2.22s. A unified total duration removes guesswork entirely.

3. Streamlines the Lucene Search Workflow

While HyperDX provides powerful SQL and Waterfall Visualization layers to look at execution streams, Lucene mode is the fastest and most standard filtering sandbox for raw text indexing. Bringing a core metric like total execution time directly into the primary search table ensures users do not have to break context or constantly hop across different Analysis Modes just to see a single metric.