thomaswyrick / xdm-helm-chart Goto Github PK

Is your feature request related to a problem? Please describe.
xDM has historically used file based logging. It is now pushing catalina logs to stdout, but they are still unstructured. However, the semarchy logs are still only written to file.

Describe the solution you'd like
Expose the catalina* and semarchy * file logs in /usr/local/tomcat/logs
Consume logs with Fluentbit and convert into json

Describe alternatives you've considered
Modifying the log4j2 xml to send to stdout and use json. The application does not respsect the xml file.

On startup have a script that provisions the repository.

Have startup probes that call the underlying xdm probes api.

On startup, have a pod that provisions the repository as the DB SA.

The Secrets are exposed at REST. Need to find a system that works easily with AKS, EKS, and potentially minikube.

FWIW, I have looked into:
External Secrets Operator
--Too complicated
Helm Secrets
--Buggy and has a client that only runs on Mac/Linux

1. Provide diagram of K8 artifacts being deployed
2. Separate K8 install instructions into separate md files.

Title says it all. Keep it tidy.

Is your feature request related to a problem? Please describe.
Users should just be able to pull down this chart using 'helm repo'

Describe the solution you'd like
Publish the chart to a public repository
Have it update automatically through a GitHub action with each new release

Describe the solution you'd like
Many xDM customers want to use K8 with their own load balancer (ex AWS ALB). Please have a HELM chart without NGINX where the active and passive services are exposed as NodePorts.

Add a third set of instructions for EKS.

Add this list:

https://www.semarchy.com/doc/semarchy-xdm/xdm/2023.1/Install/startup-configuration-reference.html

to values.yaml

Describe the bug
I cannot upload a model file.

To Reproduce
Steps to reproduce the behavior:

1. Application builder
2. New model from import
3. Select any 2023.1 model

Expected behavior
Model is imported

Kubernetes Environment
AKS

Additional context
Likely an NGINX issue

semarchy.log is not being written to stdout and so it not being picked up by the K8 logging api. The log4j2 xml looks right. Maybe this is hardcoded in xdm? Might need a sidecar log scraper (Fluentbit?).