[processor/memorylimiter] limit_percentage question

[ajsaclayan]

Hi OpenTelemetry community,

I have a question with the memory_limiter processor in regards to how it relates to container/pod resources. My team has a collector configured and deployed in k8s.

Say I have the following resources configured for a collector container:

resources:
      limits:
        memory: 1Gi
        cpu: 500m
      requests:
        memory: 500Mi
        cpu: 100m

and the following memory_limiter configured:

memory_limiter:
  check_interval: 2s
  limit_percentage: 90
  spike_limit_percentage: 20

The doc states limit_percentage is based on the following:

Maximum amount of total memory targeted to be allocated by the process heap.

How does this pertain to the requests/limits defined in this simple example?