Suggestions on what to track:

• Total number of requests coming - success v failure
• Total number of error types
• Latency/execution time
• Usage across fields

Currently, there is no prometheus instance for the beta cluster, so this needs to be product verified either locally before deployment or on production after the deployment (or both)

Metrics:

• total requests: mediawiki_wikibase_graphql_hit_total
  • has a status label with values: success, error, partial_success
• error types: mediawiki_wikibase_graphql_error_total
  • has a type label with values as listed here https://github.com/wikimedia/mediawiki-extensions-Wikibase/blob/master/repo/domains/reuse/src/Infrastructure/GraphQL/Errors/GraphQLErrorType.php ITEM_NOT_FOUND, INVALID_SEARCH_CURSOR, ...
• field usage: mediawiki_wikibase_graphql_field_usage_total
  • has a field label containing the field path
• latency: mediawiki_special_executeTiming_seconds{special="WikibaseGraphQL"}

• Alter the mediawiki_wikibase_graphql_hit_total to separate the requests coming from pure introspection vs requests querying data
• When there is an introspection only request, do not track field usage

everything else works perfectly!

In T406483#11597356, @Ifrahkhanyaree_WMDE wrote:

• Alter the mediawiki_wikibase_graphql_hit_total to separate the requests coming from pure introspection vs requests querying data
• When there is an introspection only request, do not track field usage

everything else works perfectly!

This is done now! If you make an introspection query, it will get tracked under mediawiki_wikibase_graphql_hit_total{status="introspection"}, and the fields will not be tracked.

Product verification concluded: all good.

in progress
https://grafana-rw.wikimedia.org/d/f306735c-e284-4564-a26b-9dd801f6c33c/graphql-beta?orgId=1&from=now-2d&to=now&timezone=utc

some fields are not yet available, e.g.

there is no partial success in kubernetes namespace mw-api-ext for total hits yet
there is no item not found error type in mw-api-ext yet
mediawiki_special_executeTiming_seconds{special="WikibaseGraphQL"}