[18:44:55] <jynus>	 hoo, WARNING slave_io_state Slave_IO_Running: No / CRITICAL slave_sql_lag Replication lag: 396.824618 seconds / WARNING slave_sql_state Slave_SQL_Running: No
[18:45:06] <jynus>	 that should cover your gap
[18:45:11] <jynus>	 in monitoring
[18:45:24] <hoo>	 Awesome :)
[18:45:37] <jynus>	 now we have microsecond-accurate lag measurement
[18:45:52] <jynus>	 and we can get rid of useless monitoring
[18:46:08] <jynus>	 (that works even if replica is stopped)
[18:46:37] <jynus>	 it may also help debugging lag issues
[21:07:42] <volans>	 jynus: are you taking already a look at labsdb1003?
[21:08:08] <jynus>	 yes
[21:08:47] <volans>	 what is it?
[21:14:17] <jynus>	 and there is 1.77 GiB unallocated
[21:14:27] <jynus>	 so do not lose sleep for it
[21:14:48] <jynus>	 loving LVM a bit more?
[21:14:51] <jynus>	 :-)
[21:16:32] <volans>	 lol :)
[21:18:31] <jynus>	 you know I am not a fad either, I told you, but somtimes administrative reasons are more important than yet-to-be proved performance issues
[21:18:36] <jynus>	 *fan
[21:18:54] <volans>	 ehehe I understand, btw, no mysql logs there?
[21:19:46] <jynus>	 buff, pretty much anything is disabled on labsdb, binlogs, performance schema
[21:20:05] <jynus>	 it requires every single bit we can squeeze from it
[21:21:20] <jynus>	 if you see the CPU, and iops, you will understand: https://grafana.wikimedia.org/dashboard/db/server-board?from=1455657658935&to=1455743818935&var-server=labsdb1003&var-network=eth0
[21:21:31] <volans>	 yes, saw the load