Customer Requirement:

The customer requires that OLM (Operator Lifecycle Manager) bundle unpacker jobs run with predefined resource requests and limits (CPU and memory) configured by default.

Specifically, they expect:

OLM to define reasonable default resource requests and limits for bundle unpacker jobs based on actual usage patterns, or

A supported mechanism to configure or override resource requests and limits for these jobs at cluster or namespace level.

The customer also mentioned that a common default (blanket value) across all unpacker jobs would be acceptable if appropriately sized.

Overall, while configuring the requests/limits is important, the issue presented in this ticket is they are not configured at all for OLM bundle unpacker jobs. Without having defaults set, these unpacker jobs fail in customer namespaces due to the fact that we enforce policies that require best practices.

Reason for Requirement:

In the customer environment, strict namespace policies enforce mandatory resource requests and limits as part of Kubernetes/OpenShift best practices.

Currently:

OLM bundle unpacker jobs are created without resource requests and limits.

Due to enforced policies (e.g., LimitRange, admission controllers, or policy engines), these jobs are rejected.

As a result, Operator installation fails in those namespaces.

This behavior conflicts with enterprise governance models that require all workloads to define explicit resource boundaries.