A self-contained kind-cluster setup that measures the spiffe-csi-driver's
peak memory under a worst-case scale-up burst. The scenario: a long-running
cluster suddenly schedules N CSI-using pods, kubelet drives the resulting
mount churn against the existing driver process, and the driver's peak
resident-set size grows accordingly.
The driver does not need a SPIRE control plane to exhibit this behaviour: the
allocation pattern lives entirely in pkg/mount/isMountPoint. The dummy
I hereby claim:
To claim this, I am signing this object:
Relaycorp welcomes contributions from third parties, but to protect the company and the Relaynet project from disputes and license violations, we require each contributor to sign a Contributor License Agreement (CLA).
You should sign the Relaycorp Entity Contributor License Agreement if you are making contributions on behalf of an organization or the Relaycorp Individual Contributor License Agreement if you are an individual contributor. If your organization has already signed the CLA, you should be creating PRs from forks under their GitHub organization as that makes things easier for everybody.
More information about the CLAs can be found at relaycorp.tech/cla.