Feed GraphQL feature_category field data into Error Budgets

Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.

About

This issue is part of the epic &11407. See that epic description for more context.

I #424537 we logged the performance of GraphQL fields to Kibana.

This issue is to use this data to inform error budgets for GraphQL query performance.

Proposal

We would work with groupscalability to determine how this data could inform error budgets. In &11407 (comment 1548344370) @reprazent signalled that perhaps a new form of SLI could be developed.

Our frontend records a feature category for each query it makes. This makes the group that owns that feature effectively responsible for the performance of that particular query. But queries from other clients (third-party users) are recorded with a feature category of not_owned. The scope of this new SLI (or other error budget measurement) may be just to what was previously a not_owned feature category, or it may be that this replaces the previous way of our frontend recording feature category altogether (TBC).

Edited by 🤖 GitLab Bot 🤖