Your Redis instance will evict Sidekiq data under heavy load

I have a very large site (500K pageviews/day) running on ECS and users are complaining about not getting notifications on flags. I see this in the logs:

Does this mean that I need to increase the resources of Elasticache? It’s a cache.t4g.medium and the metrics show CPU utilization in the < 5% range, so maybe sidekiq needs more RAM?

Or maybe I should just do what it says and change the maxmemory policy?

1 Like