Slightly improve logging

[noname0443]

When we use gtid diff in health check, we should ignore the gtid field on a replica, because it doesn't know the gtid of a source and always sends master state not defined.
We see now:

INFO: healthcheck: <ping=ok repl=running sync=--s ro=true offline=false lag=0.68 du=7.84% cr=no gtid=master state not defined>

I changed it to:

INFO: healthcheck: <ping=ok repl=running sync=--s ro=true offline=false lag=0.68 du=7.84% cr=no>

When we request data using getClusterStateFromDB, we know at least information about the disk usage and the daemon state on the local node and can output it to the log file.
We see now:

INFO: cs: map[mysql1:<ping=ok repl=master sync=--- ro=false offline=false lag=0.00 du=??? cr=??? gtid=gtid1:1-1820834>]
INFO: dcs cs: map[mysql1:<ping=ok repl=master sync=--- ro=false offline=false lag=0.00 du=17.40% cr=no gtid=gtid1:1-1820829>]

I changed it to:

INFO: cs: map[mysql1:<ping=ok repl=master sync=--- ro=false offline=false lag=0.00 du=10.00% cr=no gtid=gtid1:1-1820834>]
INFO: dcs cs: map[mysql1:<ping=ok repl=master sync=--- ro=false offline=false lag=0.00 du=17.40% cr=no gtid=gtid1:1-1820829>]