mgr/stats: be resilient to offline MDS rank-0
mgr/stats can repeatedly report stale perf stats when MDS rank-0 becomes offline. Even after a standby daemon transitions to active rank-0, the metrics reported are still stale.
Right now, the workaround is to wait for the query to timeout: https://github.com/ceph/ceph/blob/master/src/pybind/mgr/stats/fs/perf_stats.py#L40
To fix this, reregister user queries when a new MDS rank-0 is seen.