Probes and Health Check

Support for Kubernetes Probes

Liveness, Readiness, and Startup Probes

Kubernetes probes can terminate pods and prevent pods from receiving traffic. They're powerful tools for managing availability, but they can be dangerous if they're tuned too tightly. Review the warnings in this Kubernetes document before setting them.

Liveness, Readiness, and Startup probes are well-known methods to detect Pod health in Kubernetes. They are used in regular uptime monitoring and enable initial startup health that allows rolling deploys of new service updates.

The example below will define Liveness, Readiness, and Startup probes to one service deployment.

While creating a deployment, provide the below configuration to set up probes for your service.

Other Container Config

# setup http livenessProbe
# To perform a probe, the kubelet sends an HTTPGET request
# Service is considered healthy when /healthz retuns success code
livenessProbe:
  httpGet:
    path: /healthz #endpoint path
    port: 8080
  initialDelaySeconds: 3  #first Liveness check to be performed after 3 seconds.
  periodSeconds: 3       #verify the liveness of the application every 3 seconds

# setup http readinessProbe
readinessProbe:
  httpGet:
    path: /healthz
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 5

# setup http startupProbe
startupProbe:
  initialDelaySeconds: 1
  periodSeconds: 5
  timeoutSeconds: 1
  successThreshold: 1
  failureThreshold: 1
  exec:
    command:
      - cat
      - /etc/nginx/nginx.conf

Other Container Config

In addition to the httpGet example, TCP Probes can be configured from the Other Container Config field:

readinessProbe:
  tcpSocket:
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 10
livenessProbe:
  tcpSocket:
    port: 8080
  initialDelaySeconds: 15
  periodSeconds: 20
startupProbe:
  tcpSocket:
    port: 8080
  failureThreshold: 30
  periodSeconds: 10

Complete details of this feature are available in the Kubernetes documentation here.

Health Check

Enable Kubernetes Health by adding a Load Balancer Listener with Health Check enabled.