Investigate deployments api high redis time spent
While working on error budget improvements for ~"group::release" , we noticed GET /api/:version/projects/:id/deployments
requests are spending more time in Redis calls while those requests view serialization and cpu time spent seems to be normal.
Logs: https://log.gprd.gitlab.net/goto/f2e02b40-4e24-11ed-8d37-e9a2f393ea2a
Noticed there is no-correlation between total redis calls, redis bytes read since there are requests when the bytes read is small but the time spent in Redis is shown high.
Find the root cause of it and create follow up optimisations issue if possible.
Edited by Bala Kumar