⚠️ Hello all, I checked all documentations and discussion but couldn't find some concrete examples for "*Ory Kratos best scaling practice under heavy loads*". Here is already a discussion, https://github.com/ory/network/discussions/252, but seems we don't have quick examples. Our case is; • We have Ory Kratos hosted in our K8S • We would like to have a scalable Ory family in our cluster. • How much replicas we need to have api calls under 1 sec? (1000 customers try to login at the same time?) • Is there any configuration that we can enable/disable in https://www.ory.sh/docs/kratos/reference/configuration ? Thanks in advance for your help! This is really important for us. ❤️

the bottleneck typically is the db and not the Ory service, but you can just play with different settings, as it depends on so many factors

Thanks @steep-lamp-91158 ! You mean the database queries or we can change some configurations to increase performance..

you can scale the database obviously, but not change the queries

Is there any recommended minimum database configurations such as cpu/memory ?

depends on which one you use, with mysql it will be double of everything I guess 😅

We don't hit the DB cpu limits at all.. But somehow our developers think ORY is the bottleneck 😄 Is there a benchmark you guys did before? What is the api call limits simultaneously? (We are running in GCP, with GKE, and Postgres)