Search index health

Search index health is the operator dashboard for the queue that keeps the site's search results up to date. As repositories, issues, code, and other content change, entries are queued for indexing and processed in the background. This page shows how much work is waiting, when the indexer last ran, and how much it did, and it gives you controls to trigger indexing and to clear entries that have become stuck. The dashboard refreshes on its own every few seconds while the page is open and the tab is in the foreground.

Where to find it

URL: gitvetrix.com/admin/search
Navigation: open the admin area (your avatar menu in the top navigation bar, then Admin panel, or go to gitvetrix.com/admin). In the admin sidebar, open the Operations group and select Search.

The page heading reads Search index health.

Read the status cards

Three cards report the current state:

Queue depth — the total number of entries waiting to be indexed, with a bar breakdown by content type (such as page, issue, or code) and, when entries are waiting, the age of the Oldest pending one.
Last tick — how long ago the indexer last ran, with the count of entries it Processed and any Errors it hit on that run.
Last sweep — how long ago the last full sweep queued work, with the number of entries it Enqueued.

A timestamp that has never happened shows Never. When the queue is empty and the indexer has not yet run, the page shows No search indexing activity yet — queue empty. instead of the cards. If the status cannot be loaded, a banner reads Search indexing temporarily unavailable. Retrying... and the page keeps retrying on its own.

Trigger a full sweep

Select Trigger full sweep to queue every eligible item for reindexing. A confirmation message reports how many entries were added and the new queue depth. Use this after a change that should refresh the whole index.

Clear stuck entries

Two controls handle entries that stop making progress. They address different problems, so use the one that matches what the dashboard reports.

Reset stranded claims — releases entries that were picked up for indexing but never finished, so the indexer can pick them up again on its next run. A confirmation message reports how many were released.
Reset retry-budget-exceeded rows — clears entries that have failed enough times to use up their retry budget so the indexer tries them again. When any such entries exist, the Queue depth card shows a warning that counts them and points you to this control. Selecting it opens a confirmation dialog where you set how many entries to clear and the failure threshold; select Reset rows to confirm. A confirmation message reports how many were cleared.

Resetting stranded claims does not clear retry-budget-exceeded entries, and resetting retry-budget-exceeded entries does not release stranded claims.

Audit Log — the instance-wide record of administrative events.
Analytics — instance-wide usage metrics and leaderboards.
Administration — the full list of admin pages.