Get starting screen when migrating to 0.40.0


Our current setup is to run the metabase application in ECS in a container, and we have a RDS Postgres to which we connect to from the container, until now we never had issue upgrading Metabase but now when migrating to v0.40.0 the metabase application shows the "starting screen".

Any idea what the issue could be ?

Thanks in advance,


Hi @alex23
So you are not using Elastic Beanstalk coupled with RDS? If you are using EB, then you need to decouple it from RDS:
If you are just using a plain Docker, then your MB_DB_* environment variables are not applied.

No indeed I'm using plain docker, where can I find information about those MB_DB_* environment variables as I believe I have never had to set those variables?


@alex23 If you have never set the environment variables, then you were using the H2 internal database in the container - which means that is deleted, when you destroyed the container. So you should revert to your backup.
I don't understand if you were using RDS for the application database or as a data source.

I use RDS for the application database and the current setup uses those env variables to connect to the RDS database: RDS_DB_NAME, RDS_USERNAME, RDS_HOSTNAME, RDS_PORT, RDS_PASSWORD but apparently the naming of those has been changed.

I have checked the log to verify the application database:
"application-database": "postgres",
"application-database-details": {
"database": {
"name": "PostgreSQL",
"version": "11.10"
"jdbc-driver": {
"name": "PostgreSQL JDBC Driver",
"version": "42.2.18"

@alex23 Use the MB_DB_* environment variables instead of RDS_*

@alex23, I am trying to get my decoupled rds to connect to the new aws beanstalk sans-rds.. their instructions are not the clearest. its more of a decouple this, and figure out the rest for yourself, as their step by step guide just keeps looping on itself. I have slowly be documenting the steps, hoping to have it finished today, and I will update this.