Describe the bug

When installing langfuse for the first time using Helm, the langfuse-web container gets stuck in a crash-loop.

What's actually happening is:

1. the web container starts up
2. it runs the entrypoint.sh script
3. the script attempts to run prisma migrations
4. in the meantime kubernetes is checking for health
5. since the migrations take a long time on a new install the health check fails
6. web pod is terminated.
7. migration is left in a failed state
8. web restarts, retried the entrypoint script.
9. Can't complete migrations
10. Rinse and repeat ad-infinitum

The best solution to this problem is to implement an intialization container that runs before the 'main' web container starts.

To reproduce

Just do a basic helm install with the default values proposed in the documentation.

Chart version and values.yaml configuration

Version 1.5.9

langfuse:
image:
repository: langfuse/langfuse
tag: "3.131"
salt:
value: "secretSalt"
nextauth:
secret:
value: "secretSecret"

postgresql:
auth:
password: "secretPassword"

clickhouse:
auth:
password: "secretPassword"

redis:
auth:
password: "secretPassword"

Kubernetes Status

No response

Logs

No response

Additional information

No response

Are you interested to contribute a fix for this bug?

Yes