grafana
[Docs] Add sizing recommendation for backend pod
Describe the solution you'd like The documentation about "size the cluster" could add support for the new SSD chart with the three part deployment. The current version as of now only show write/read. Example:
loki:
authenabled: false
read:
replicas: 75
resources:
requests:
cpu: 3
memory: 6
limits:
cpu: 3
memory: 8
write:
replicas: 58
resources:
requests:
cpu: 1
memory: 6
limits:
cpu: 2
memory: 12
Describe alternatives you've considered N/A
Additional context https://grafana.com/docs/loki/next/installation/sizing/
Please also note that unit must be specified for the memory. Without any, the default will be taken into account. The default unit for memory is byte and we don't want to start with that few memory.
This would be really helpful!
The docs also don't mention any sizing recommendations.
+1
I think such recommendations would be good for each deployment mode. The docs suggest certain ingestion rates, would be great to have docs for the resource requests / limits. Maybe even adding sane defaults to the helm chart? What do you guys think about that?
+1
The sizing recommendations provide excellent guidance for deploying Loki on Kubernetes in SSD mode, especially as a solid starting point. These recommendations can be adjusted and scaled as needed later. If the sizing recommendations are not used, it's essential to at least define reasonable resource requests and limits in the Helm values to ensure a reliable initial setup.
