feat(app): histogram and filters across every selected source - #2886
feat(app): histogram and filters across every selected source#2886teeohhem wants to merge 1 commit into
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
🦋 Changeset detectedLatest commit: 7a9363b The changes in this PR will be included in the next version bump. This PR includes changesets to release 3 packages
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
Greptile SummaryThis PR completes multi-source search by extending the histogram, total count, filter facets, value distributions, and pinned filters across all selected sources.
Confidence Score: 3/5The PR does not appear safe to merge until duplicate source names retain distinct histogram series and partial source-query failures are disclosed rather than shown as complete aggregates. The current histogram groups and colors sources by non-unique display name, while its merge path silently omits failed source responses and still renders the remaining histogram and count as a complete result. Files Needing Attention: packages/app/src/components/SearchHistogram.tsx
|
| Filename | Overview |
|---|---|
| packages/app/src/DBSearchPage.tsx | Connects per-source search specifications to the histogram, total count, and filters sidebar while preserving the single-source rendering path. |
| packages/app/src/components/SearchHistogram.tsx | Implements per-source histogram querying and aggregation, but duplicate source names still collide and partial query failures still produce undisclosed incomplete aggregates. |
| packages/app/src/components/DBSearchPageFilters.tsx | Changes the filter sidebar to consume multiple source configurations and merge their facets, distributions, columns, and pins. |
| packages/app/src/components/DBSearchPageFilters/hooks.ts | Adds slot-based per-source facet querying and unions the returned fields and values. |
| packages/app/src/hooks/useMetadata.tsx | Adds per-source value-distribution queries and sums counts by value across successful sources. |
| packages/app/src/searchFilters.tsx | Extends pinned-filter state and mutations across the selected source IDs. |
Flowchart
%%{init: {'theme': 'neutral'}}%%
flowchart LR
Sources[Selected sources] --> Specs[Per-source search specifications]
Specs --> Queries[Per-source ClickHouse queries]
Queries --> Histogram[Stacked multi-source histogram]
Queries --> Count[Summed result count]
Specs --> Facets[Merged facet fields and values]
Specs --> Distributions[Summed value distributions]
Sources --> Pins[Union and fan-out pinned filters]
Facets --> Sidebar[Filters sidebar]
Distributions --> Sidebar
Pins --> Sidebar
Reviews (2): Last reviewed commit: "feat(app): histogram and filters across ..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| const sourceName = specs[i].source.name; | ||
| for (const row of response.data ?? []) { | ||
| data.push({ ...row, [SOURCE_GROUP_COLUMN]: sourceName }); |
There was a problem hiding this comment.
When two selected sources have the same display name, both responses receive the same __hdx_source group value, causing the chart transform to collapse them into one series and the name-keyed color map to retain only one source color.
Knowledge Base Used: App Components and Charts
| const isLoading = slots.some(s => s.isLoading); | ||
| const allFailed = slots.length > 0 && slots.every(s => s.isError); | ||
| const anyError = slots.some(s => s.isError); | ||
| const error = slots.find(s => s.error != null)?.error ?? undefined; | ||
| const isComplete = | ||
| slots.length > 0 && slots.every(s => s.isError || !!s.data?.isComplete); |
There was a problem hiding this comment.
When one source query fails while another succeeds, the merge omits the failed source and only renders an error when every source fails, causing the histogram and total count to present an incomplete aggregate without disclosing it.
Knowledge Base Used: App Components and Charts
| const mergedResponse: ResponseJSON<Record<string, any>> | undefined = | ||
| useMemo(() => { | ||
| let meta: ColumnMetaType[] | undefined; | ||
| const data: Record<string, any>[] = []; |
There was a problem hiding this comment.
The new 421-line component combines querying, aggregation, formatting, and rendering while using any for response rows; NestedFilterGroup also introduces any[] for chart configurations. This bypasses the repository's proper-typing and 300-line component requirements, reducing compile-time protection and making the feature harder to maintain.
Context Used: CLAUDE.md (source)
Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!
The results table already spanned sources; the histogram and the filters sidebar still picked a component per case at the call site. The page now renders SearchHistogram and SearchTotalCount, which take the same per-source list the table does. One source keeps the full DBTimeChart — severity grouping, series drill-down and focus, the pinned tooltip — because a single source has a severity vocabulary to group by. Several sources can't share one, so they stack a count() series each. DBSearchPageFilters takes a list too, so the sidebar people already use works across a selection: facet fields and values merge, value counts are summed so a percentage describes the whole search, "load more" fans out and unions, and pins read as a union and write to every selected source. A source whose table lacks a filtered column is excluded from the results with the reason on its status chip, rather than quietly returning rows that ignore the filter. With one source every path is the one it was before: analysis-mode tabs, denoise, shared filters, and dropping percentages when a distribution query fails.
c6e26b2 to
7a9363b
Compare
E2E Test Results✅ All tests passed • 279 passed • 1 skipped • 1153s
Tests ran across 4 shards in parallel. |
1 participant
Stacked on #2885. Completes multi-source search.
#2885 made the results table N-ary but left the histogram and filters sidebar single-source. This finishes both, so multi-source is fully featured rather than a reduced mode.
Histogram and count. The page renders
SearchHistogram/SearchTotalCount, which take the same per-source list the table does. One source keeps the fullDBTimeChart— severity grouping, series drill-down and focus, the pinned tooltip — because a single source has a severity vocabulary to group by. Several sources can't share one, so they stack acount()series each.Filters sidebar.
DBSearchPageFilterstakes a list too, so the sidebar people already use works across a selection: facet fields and values merge, value counts are summed so a percentage describes the whole search, "load more" fans out and unions, and pins (personal and team-shared) read as a union and write to every selected source — pinServiceNamewhile searching logs and traces and it stays pinned in both. A source whose table lacks a filtered column is excluded with the reason on its status chip instead of quietly returning rows that ignore the filter.With one source every path is the one it was before: analysis-mode tabs, denoise, shared filters, and dropping percentages when a distribution query fails.
Verified live against the demo stack before this stack was split (the split is provably behavior-preserving: this branch's production code is byte-identical to the verified tree). Single source keeps its tabs, denoise, counts, pins, "Show more" and "More filters"; two sources show 33 merged facet groups including nested
LogAttributesandSpanAttributesside by side, logs-onlySeverityTextnext to traces-onlyStatusCode, summed counts, and pins.