Page MenuHomePhabricator

Add Link: Set up cronjob for collecting statsd metrics about dangling search index entries
Closed, ResolvedPublic

Description

Set up a cronjob to monitor what fraction of the tasks in the search index are missing from the database (ie. would result in a "no suggestions" error when clicked by the user), and create a grafana dashboard.

fixLinkRecommendationData.php with the --dry-run --statsd option can almost do this, except search can only return 10K results. So either use the --random flag (which is in wmf.20) or make it able to iterate through more search results (T284531: Add Link: Work around 10K search result set limit in fixLinkRecommendationData.php).

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

We might want to collect stats about dangling DB entries as well while we are at it. They are much less problematic but it's no extra effort.