In large clusters, a surge of JMX reconnection can generate a very high load of QUORUM updates on the cluster table just to update the last contact timestamp.
We can lower our CL to LOCAL_ONE which will lighten the footprint, especially on large multi DC setups.
In large clusters, a surge of JMX reconnection can generate a very high load of QUORUM updates on the cluster table just to update the last contact timestamp. We can lower our CL to LOCAL_ONE which will lighten the footprint, especially on large multi DC setups.