How big is your concourse deployment.

[skreddy6673]

I'm just starting this discussion to see who big the concourse deployments are. This may help some individuals to decide the amount of resources do they need to allocate to their deployment to make sure it's stable.

I'm starting with mine:

Helm deployment: concourse: v5.7.1; Helm Chart: v8.2.13 
Web nodes: 3 -  each with 7 CPU and 30Gi Memory
Worker nodes: 12 - each with 8 CPU, 60Gi Memory and 600Gi storage
DB: PostgreSQL - db.m4.xlarge - 229 GiB storage
The Team and Pipeline count doesn't matter but still including 
1300+ pipelines spread across 188 teams. 

Fell free to ignore this if you don't find value in this.
Thanks

[skreddy6673]

Is there any way to see how much resources we should be allocating to concourse?

Helm deployment: concourse: v5.7.1; Helm Chart: v8.2.13
Web nodes: 3 - each with 7 CPU and 40Gi Memory
Worker nodes: 12 - each with 15 CPU, 60Gi Memory and 600Gi storage
DB: PostgreSQL - db.m5.4xlarge
The Team and Pipeline count doesn't matter but still including
1300+ pipelines spread across 188 teams.

I still have issues where kubernetes kill workers with OOM message. Is increasing the worker resources my only chance?