Hi Andrew,
I noticed the connection error you posted on the GitHub issue has a 3000 ms timeout:
com.mongodb.MongoTimeoutException: Timed out after 3000 ms
The normal driver default for
server selection timeout is 30,000ms (30s). Where is your Metabase application deployed in relation to your Atlas cluster -- are both hosted by the same cloud provider in the same region?
I'm not familiar with Metabase or Monger configuration, but if your connection error is correct I would look into changing the server selection timeout to the default in case 3 seconds is too short. Based on your GitHub issue it looks like you are able to use standard MongoDB connection strings so I would try appending
&serverSelectionTimeoutMS=30000.
Have you checked your MongoDB Atlas logs to see if there is any relevant information logged about connection attempts? To simplify troubleshooting I would remove the readPreference from your connection string so attempts will be logged on the primary.
If adjusting the server selection timeout doesn't work, can you provide your specific versions of Metabase, Monger, and JDK in case those provide some useful hints? You mention this looks like Monger v3.5.0 based on a git commit, but it would be helpful to confirm what is actually installed in your environment.
There is also a
Metabase discussion forum, which I expect would have users with more expertise on Metabase configuration. If you haven't already posted there, it would be worth starting a discussion (and adding the link here so users with a similar issue can follow the breadcrumb trail).
Regards,
Stennie