Completed a 23.1 to 23.3 on-premise upgrade (up to HF03). Most things work but having several issues that all seem to potentially be networking- and/or pod/service communication- related. A couple of the settings panels are returning a 404 nginx error page. Consistently getting "data fetch" errors from things like /oi/v2/... url's at the bottom right corner, and we're seeing a weird error trying to communicate to Remedy from the incident management pod. These might be all independent and/or they might be related to a common network config root cause.
Resolution
The problem is that the default value for NOTIFY_FILTER_PORT inside the pod was incorrect. The value was in the format tcp://xxx.xxx.xxx.xxx:8090 where it should only be the port number
The resolution was to add an environment variable manually to the incident management deployment as follows:
NOTIFY_FILTER_PORT: "8090"
As of now, it seems we do not require any further changes to the configuration.
Additional Information
NOTIFY_FILTER_PORT env variable default value is now configured within the image itself. Looks like Kubernetes is creating default env variables for each pod. Following are the examples collected from the 24.1 environment. In this case, the Kubernetes default env variable name, the env variable name that we have defined in incident management for NOTIFY_FILTER_PORT matched.
Because of the default env variable from Kubernetes, we got a different value than what we expected, which is causing the issue here.