#8 Configure default quotas on CentOS CI Infra OCP4 projects
Closed: Duplicate 2 years ago by zlopez. Opened 3 years ago by dkirwan.

We currently do not have quotas/limits on the resources available to tenants on the CentOS CI Infra OCP4 clusters.

We need to spend a little time investigating and implementing a default quota/limit for tenants.

eg: 2vCPU, 2GB ram, 1 kubevirt VM with similar resources


no_quota_wild_west.png

Sharing the usage of a single tenant on the OCP4 production cluster last night:

  • 38 pods
  • 97.28 CPU cores
  • 138.6 GB RAM

We must set quotas on our tenants usage to prevent this kind of behaviour.

I propose we create a few tiers, default (small), medium, large quota levels, maybe something like:

Default (Small)
- 5 pods
- 2 CPU cores
- 4 GB RAM
- 1 Kubevirt VM

Medium
- 10 pods
- 5 CPU cores
- 15 GB RAM
- 2 Kubevirt VM

Large
- 15 pods
- 10 CPU cores
- 30 GB RAM
- 5 Kubevirt VM

Large Memory Intensive Workloads
- 15 pods
- 10 CPU cores
- 50 GB RAM
- 5 Kubevirt VM

Large CPU Intensive Workloads
- 15 pods
- 15 CPU cores
- 30 GB RAM
- 5 Kubevirt VM

Even our large tenants should be able to fit comfortably within the large tiers unless they are misbehaving!

Metadata Update from @dkirwan:
- Issue tagged with: groomed

3 years ago

Metadata Update from @dkirwan:
- Issue priority set to: Waiting on Assignee

3 years ago

As suggested by @siddharthvipul1 via email, I created https://pagure.io/centos-infra/issue/34 for configuring a default LimitRange on projects.

Metadata Update from @dkirwan:
- Issue marked as depending on: #34

3 years ago

By default, each namespace will have 4 GB ram quota and no PVs quote since we already control it with NFS.
(discussed in the meeting and mutually agreed)

@dkirwan Do you want to work on this ticket?

Metadata Update from @dkirwan:
- Issue assigned to dkirwan

2 years ago

[backlog refinement]
@dkirwan Did you have time to look at this?

Metadata Update from @zlopez:
- Issue tagged with: blocked

2 years ago

[backlog refinement]
It was identified as part of OpenShift optimization.

[backlog refinement]
It's part of the upcoming replacement of OpenShift in AWS.

Metadata Update from @zlopez:
- Issue close_status updated to: Duplicate
- Issue status updated to: Closed (was: Open)

2 years ago

Login to comment on this ticket.

Boards 1
CentOS CI Infra Status: Triaged
Attachments 1
Attached 3 years ago View Comment