Reduce noise on Mesos reconciliation.

This log line has negative performance impact at scale (thousands of
tasks) as the scheduler ends up having high long tail scheduling
iteration latencies. The problem gets worse depending on the scheduling
interval, for example, every 5 seconds.

Trying this out in production yields consistent double-digit millisecond
latencies as opposed to single second peaks from before.

server/src/main/java/io/mantisrx/server/master/mesos/MesosSchedulerCallbackHandler.java

@@ -332,7 +332,7 @@ public void slaveLost(SchedulerDriver arg0, SlaveID arg1) {
     @Override
     public void statusUpdate(final SchedulerDriver arg0, TaskStatus arg1) {
         Optional<WorkerId> workerIdO = WorkerId.fromId(arg1.getTaskId().getValue());
-        logger.info("Task status update: ({}) state: {}({}) - {}",
+        logger.debug("Task status update: ({}) state: {}({}) - {}",
                 arg1.getTaskId().getValue(),
                 arg1.getState(),
                 arg1.getState().getNumber(),