Memory requests and limits¶

Recommendations¶

The values need regular observation to see the baseline memory usage and spikes
Always use memory limits
If you want overprovisioning, add memory requests to a level slightly above the average baseline memory usage and add memory limits to absorbe spikes.
If not, always set your memory requests equal to your limits to a level to absorbe spikes.
Use Horizontal pod autoescaler for workloads with replicas
Investigate to use vertical pod autoescaling in workloads without replicas

Example:

What Everyone Should Know About Kubernetes Memory Limits, OOMKilled Pods, and Pizza Parties