Mayne0213
4286296591
- CPU throttling prevents app startup, not crashes - Memory OOM is the real cascading failure cause - CPU request ensures fair scheduling
28 lines
732 B
YAML
28 lines
732 B
YAML
# VPA Helm Values
|
|
# Reference: https://github.com/FairwindsOps/charts/tree/master/stable/vpa
|
|
|
|
# Recommender - recommends resource requests based on historical usage
|
|
recommender:
|
|
enabled: true
|
|
replicaCount: 1
|
|
|
|
resources:
|
|
requests:
|
|
cpu: 15m
|
|
memory: 128Mi
|
|
limits:
|
|
memory: 192Mi
|
|
|
|
# Updater - applies recommended resource requests to pods
|
|
# Disabled because we're using updateMode: Off (recommendations only)
|
|
updater:
|
|
enabled: false
|
|
|
|
# Admission Controller - injects resource requests into new pods
|
|
# Disabled because we're using updateMode: Off (recommendations only)
|
|
admissionController:
|
|
enabled: false
|
|
|
|
# Optional: Configure default VPA behavior
|
|
# metrics-server should be installed in the cluster
|