In order to verify that charm reflects the workload's health status, you can deploy the charm configured with a dummy URL that is not reachable. This should result in the workload not being healthy and that should be reflected in the charm's status once an update-status event is fired.
juju deploy oidc-gatekeeper --channel latest/edge/pr-134 --trust --base ubuntu@20.04 --config public-url=http://wrong-url
### wait 40 seconds in order for workload health check to go down `(threshold * (period + timeout))`
sleep 40
jhack fire oidc-gatekeeper/0 update-status
juju status
Implementation
This PR adds:
Maintenance: Workload failed health check
if workload has hit health checks failure thresholdlatest/edge
dex-auth
to be reachableActive
once it is configured with properpublic_url
For implementation decisions about the health check and tests, please take a look at the issues being closed.
This PR uses for testing PR https://github.com/canonical/charmed-kubeflow-chisme/pull/88 and for the charm PR https://github.com/canonical/charmed-kubeflow-chisme/pull/85. They should be merged to main and either release new chisme and unpin chisme or pin to
main
branch in bothrequirements*.in
files. Otherwise, this will break once we merge and branches are deleted.Testing
In order to verify that charm reflects the workload's health status, you can deploy the charm configured with a dummy URL that is not reachable. This should result in the workload not being healthy and that should be reflected in the charm's status once an
update-status
event is fired.Closes #82 Closes #130