Reputation: 2033
I have a pod which keeps restarting because of failed liveliness probes:
Events:
... Container ... failed liveness probe, will be restarted
I suspect the the liveliness timeout of 1 sec is the issue here.
The liveliness probe is defined as follows:
livenessProbe:
httpGet:
path: /health
port: 80
scheme: HTTP
Is there a way to simulate the HTTP request with kubectl
to take a few samples on the response time?
exec
ing into the container and running curl
is not an option, because the container is distroless.
Felix
Upvotes: 1
Views: 603
Reputation: 2504
You can port-forward to the container with kubectl and then try to query the health api endpoint.
Also, slow down your livenessProbe
while you check that with initialDelaySeconds: 120
(in the livenessProbe
object, outside the httpGet
)
Like so:
livenessProbe:
initialDelaySeconds: 120
httpGet:
path: /health
port: 80
scheme: HTTP
Upvotes: 2