Kalix services redeploy CLI command does not find anything to redeploy

Hi,

We’ve converted most of our akka serverless services to Kalix and deployed them last Friday, did some basic endpoint tests and everything was working correctly.

Today, all those services respond with a 503/no healthy upstream. When looking in the dashboard or via the CLI, the services are listed as Ready, but on the dashboard detail page, the service instance status is Pending.

I tried redeploying the services. This is the output of the kalix services list command:

NAME                      AGE     REPLICAS   STATUS        DESCRIPTION   
asset-type-service        4d20h   1          Ready                       
legacy-acl-service        4d19h   1          Ready                       
task-dispatcher-service   4d19h   1          Ready 

I then execute the kalix services redeploy task-dispatcher-service command, and this is the output:

There is nothing to redeploy for service: task-dispatcher-service
Are you sure this name is correct?

I noticed some errors in the logs pertaining to the logback.xml (I forgot to change the appender layout class from Akka Serverless to the new Kalix class :man_facepalming:), so I created a new image & tag. I tried executing kalix services deploy task-dispatcher-service my-registry/task-dispatcher-service:new-tag with following output:

Service 'task-dispatcher-service' was successfully updated to generation 3.

Nothing seems to be happening however. Any suggestions on what to do other than maybe deleting the service and deploying it again? Thanks in advance.

Hi Guillaume,

There is a known problem with redeploy and fix is in progress. It will be renamed to restart, because it is just about restarting the service, not deploying a new version.

deploy with a new image tag should update the service

We can continue the troubleshooting in our Slack channel.

1 Like

Hello Patrik,

Sorry for the late reply, I had a day off. Apparently a little time off was all that was needed because the services seem to have recovered in the meantime without any actions on our part, so no need for further troubleshooting. But it is good to know that a fix regarding the redeploy/restart is in the works :slightly_smiling_face: