subscriber-config not populated using suggested litmus-agent inputs

[kbmulleady]

When installing the litmus-agent helm chart, I use the following suggested values from the README for a self agent in cluster mode:

LITMUS_USERNAME: "admin"
LITMUS_PASSWORD: "litmus"
LITMUS_ENVIRONMENT_ID: "sandbox"
LITMUS_PROJECT_ID: "<pulled_from_chaos_portal>"
LITMUS_URL: "http://litmus-frontend-service.litmus.svc.cluster.local:9091"
LITMUS_BACKEND_URL: "http://litmus-server-service.litmus.svc.cluster.local:9002"
APP_VERSION: "3.11.2"
global:
  INFRA_MODE: "cluster"

However, with these values the subscriber-secret and subscriber-config aren't populated. The event-tracker and subscriber pods also complain that the VERSION and ACCESS_KEY values are missing respectively.

From looking at some suggestions in the Slack channel, I tried setting the LITMUS_BACKEND_URL to http://litmus-server-service.litmus.svc.cluster.local:9002/query which caused the subscriber-config and subscriber-secret to be populated but still incorrect. With this setup, the SERVER_ADDR value in the subscriber-config is set to http://litmus-server-service.litmus.svc.cluster.local:9002/query/query. This causes the subscriber to get the error message:

level=fatal msg="Failed to parse agent confirm data" data="404 page not found" error="invalid character 'p' after top-level value"

If I patch the subscriber-config to set the SERVER_ADDR to http://litmus-server-service.litmus.svc.cluster.local:9002/query, the subscriber starts running but has the error message:

level=error msg="Error response from the server : {\"payload\":{\"errors\":[{\"message\":\"ERROR: accessID MISMATCH\",\"path\":[\"infraConnect\"]}],\"data\":null},\"type\":\"data\"}\n"

I checked to make sure the INFRA_ID value in the subscriber-secret matched the infra ID in the chaos portal and they do match.

I have been fighting with this deployment for quite some time with no luck. Any help would be greatly appreciated, thanks!

[kbmulleady]

This ended up being an issue on my end. I had set the graphqlServer and quthServer services to be of type NodePort which caused these connectivity issues on startup. Removing those settings and using the default LITMUS_URL and LITMUS_BACKEND_URL values as described in the README worked.