Is it possible to define a custom liveness/readiness probe for a integration server custom resource deployed on Openshift?


Currently, the spawned integration server deployment/pod is using a hard-coded check 'chkacehealthy':

livenessProbe:
exec:
command:
- chkacehealthy

- IBM-AppConnectEnterprise-Openshift/main.go at master · ivancamargo190/IBM-AppConnectEnterprise-Openshift

Is it possible to overwrite this value with a different command, that is, to ping a different endpoint? The integration server CR only mentions probe time thresholds, but no possibility of a different probe check. And changing the created deployment resource is a risky take as there is no guarantee it will persist, and not be changed by the ACE operator.


Our goal is to have a liveness probe that will confirm that a deployed integration server has a running integration flow. The motivation is due to previous experience where our IIB env had situations where the integration node/server would be reported as UP, but the flows would not be responding, which required us to restart the server. If we are able to probe a dummy flow, its unresponsiveness would trigger a pod recreation.

------------------------------
Vid Romac
------------------------------

Hi @Vid Romac

Apologies for such a late response - here is what our team said about this: 

If you look here at the Custom resource values: https://www.ibm.com/docs/en/app-connect/containers_cd?topic=resources-integration-server-reference  The runtime is the App Connect integrationServer runtime container. eg spec.pod.containers.runtime.livenessProbe.timeoutSeconds

Let me know if this helps and again, apologies! 

------------------------------
Gabriel Marte Blanco
Austin TX
------------------------------

Original Message:
Sent: Fri January 27, 2023 04:53 AM
From: Vid Romac
Subject: Custom probes for ACE containers, integration server

Is it possible to define a custom liveness/readiness probe for a integration server custom resource deployed on Openshift?


Currently, the spawned integration server deployment/pod is using a hard-coded check 'chkacehealthy':

livenessProbe:
exec:
command:
- chkacehealthy

- IBM-AppConnectEnterprise-Openshift/main.go at master · ivancamargo190/IBM-AppConnectEnterprise-Openshift

Is it possible to overwrite this value with a different command, that is, to ping a different endpoint? The integration server CR only mentions probe time thresholds, but no possibility of a different probe check. And changing the created deployment resource is a risky take as there is no guarantee it will persist, and not be changed by the ACE operator.


Our goal is to have a liveness probe that will confirm that a deployed integration server has a running integration flow. The motivation is due to previous experience where our IIB env had situations where the integration node/server would be reported as UP, but the flows would not be responding, which required us to restart the server. If we are able to probe a dummy flow, its unresponsiveness would trigger a pod recreation.

------------------------------
Vid Romac
------------------------------

I am aware of the properties exposed via the integration server CR, however the values exposed are primarly focused on the probe period/delay/threshold values; there is no way to customize the probe itself: which endpoint to call or a custom behaviour how to call it.

We solved this with a custom OCP CronJob that creates a pod containing our script that goes thru a list of services to check and the endpoints they expose. All services that don't respond, have their pods killed, which triggers new ones to spawn.

------------------------------
Vid Romac
------------------------------

Original Message:
Sent: Fri February 10, 2023 10:50 AM
From: Gabriel Marte Blanco
Subject: Custom probes for ACE containers, integration server

Hi @Vid Romac

Apologies for such a late response - here is what our team said about this: 

If you look here at the Custom resource values: https://www.ibm.com/docs/en/app-connect/containers_cd?topic=resources-integration-server-reference  The runtime is the App Connect integrationServer runtime container. eg spec.pod.containers.runtime.livenessProbe.timeoutSeconds

Let me know if this helps and again, apologies! 

------------------------------
Gabriel Marte Blanco
Austin TX

Original Message:
Sent: Fri January 27, 2023 04:53 AM
From: Vid Romac
Subject: Custom probes for ACE containers, integration server

Is it possible to define a custom liveness/readiness probe for a integration server custom resource deployed on Openshift?


Currently, the spawned integration server deployment/pod is using a hard-coded check 'chkacehealthy':

livenessProbe:
exec:
command:
- chkacehealthy

- IBM-AppConnectEnterprise-Openshift/main.go at master · ivancamargo190/IBM-AppConnectEnterprise-Openshift

Is it possible to overwrite this value with a different command, that is, to ping a different endpoint? The integration server CR only mentions probe time thresholds, but no possibility of a different probe check. And changing the created deployment resource is a risky take as there is no guarantee it will persist, and not be changed by the ACE operator.


Our goal is to have a liveness probe that will confirm that a deployed integration server has a running integration flow. The motivation is due to previous experience where our IIB env had situations where the integration node/server would be reported as UP, but the flows would not be responding, which required us to restart the server. If we are able to probe a dummy flow, its unresponsiveness would trigger a pod recreation.

------------------------------
Vid Romac
------------------------------