I’m not sure what happened, but when I upgraded my Atlas cluster from an M0 to an M10, suddenly the number of connections quickly shot up to 1500. When I remove all IP addresses from Network Access, the lowest it would go down to is about 1200. Where are these connections coming from and how do I troubleshoot it?
Is there something in the dedicated cluster configuration that I’m not considering
Dedicated clusters do have a higher maximum connection count but include internal connections between the node(s) and also other atlas monitoring or automation agents. However, I don’t believe this should reach 1200 as it generally only counts for approximately 30-60 connections (out of a maximum of 1500 for M10 tier clusters).
In comparison to the shared tier clusters, the internal atlas connections do not get counted as a part of the connection count. e.g. A newly created M0 with no external connections should be at 0 connections in the Atlas metrics where as an M10+ tier cluster with no external connections would normally be expected to be at 30-60 connection count.
When I remove all IP addresses from Network Access, the lowest it would go down to is about 1200.
Where are these connections coming from and how do I troubleshoot it?
Even after some time has passed after removing all IP addresses from your Network Access list, you can open a support chat/case to verify what connections are being made to the cluster if you have 0 IP’s in the access list. Additionally, you can try view and download the MongoDB logs to see if there are any particular IP’s that are connecting the most often.