CPU LIMIT INCREASE

Question

Hi Team,

I have a question regarding the ARO quota.

When I check my quota using the command oc get quota, the CPU limit shows as 5800m/6. However, when I sum up the CPU limits assigned to each pod, the total comes to 4750m.

Could you please help me understand the reason for this difference?

Answer 1

Hello,

Welcome to Microsoft Q&A,

The extra ~1050m you’re seeing (5800m shown in the quota vs. 4750m summed by hand) almost always comes from things your manual sum didn’t include:

• Init containers — quota counts the max CPU limit across all init containers per pod (in addition to app containers).
• Pod Overhead — if your RuntimeClass sets pod overhead, that CPU is added to the pod’s accounted limits.
• Defaulted limits from a LimitRange — containers without explicit limits get a default limit applied, which counts toward quota even if you didn’t specify it.
• Terminating / pending / failed-but-not-garbage-collected pods — ResourceQuota usage is decremented only after the pod object is actually deleted. These can linger and still count.
• Ephemeral OpenShift pods — build/deployer hooks, Jenkins agents, etc., in the same namespace also count while present.
• Scoped quotas — if your project has multiple ResourceQuota objects (e.g., NotBestEffort vs Terminating scopes), the used you’re looking at may represent a subset that’s larger than your hand-picked pod list.

See exactly what the quota is counting

oc describe quota <quota-name>

oc get limitrange
oc describe limitrange
oc get limitrange -o yaml

Check:

scopes: (e.g., NotBestEffort, Terminating)

• the limits.cpu used and hard

whether there are multiple quota objects

Please upvote and accept the answer if it helps!!