We are experiencing random MongoDB freezes on our 2 replicasets. During short time intervals (about 30-60 seconds) we are unable to get response from DB. That happens randomly. No CPU, Disk, Network or memory spikes are observed in our Nagios monitoring.
MongoDB logs show that db accepts connections and after 30 seconds client disconnects.
From the application side a bunch of MongoCursorTimeoutException's were thrown from a PHP driver at March 21, 2018, 3:01:42 a.m. UTC.
I've attached diagnostic.data logs for the problematic period of time from our 3 db instances.
DB1 - master, DB2/DB3 - slaves. The problem started on DB2 at March 21, 2018, 3:01:12 and 30 seconds later we've observed exceptions on our PHP application servers.
Digging into diagnostic logs I've found that the problematic period of time is empty there.
Is this problem resolved in a latter MongoDB releases? We are planing an update and want to make sure that the fix is already included.