Change Details

https://icinga.wikimedia.org/cgi-bin/icinga/status.cgi?search_string=tendril ^ the green one is OK, but the yellow ones should be gone This is because the monitoring check gets added per role::tendril, and moving tendril from einsteinium to dbmonitor is ongoing thing afaict. Multiple hosts get the check but only one actually has the tendril name in DNS, as of right now that is still einsteinium. So there is no real cert expiry issue here, the actual tendril.wm.org is fine and has LE autorenewal. Add some setting to get the monitoring just on the host that currently has the tendril CNAME in DNS whichever it is, so that the role can be moved freely without causing false alerts.

https://icinga.wikimedia.org/cgi-bin/icinga/status.cgi?search_string=tendril ^ the green one is OK, but the yellow ones should be gone ``` dbmonitor1001 HTTPS-tendril This service problem has been acknowledged WARNING 2017-04-04 18:56:41 1d 17h 2m 52s 3/3 SSL WARNING - Certificate tendril.wikimedia.org valid until 2017-04-10 01:53:01 +0000 (expires in 5 days) dbmonitor2001 HTTPS-tendril This service problem has been acknowledged WARNING 2017-04-04 18:55:54 1d 17h 8m 40s 3/3 SSL WARNING - Certificate tendril.wikimedia.org valid until 2017-04-10 01:47:36 +0000 (expires in 5 days) einsteinium HTTPS-tendril OK 2017-04-04 18:56:42 54d 4h 48m 26s 1/3 SSL OK - Certificate tendril.wikimedia.org valid until 2017-05-28 03:56:00 +0000 (expires in 53 days) ``` This is because the monitoring check gets added per role::tendril, and moving tendril from einsteinium to dbmonitor is ongoing thing afaict. Multiple hosts get the check but only one actually has the tendril name in DNS, as of right now that is still einsteinium. So there is no real cert expiry issue here, the actual tendril.wm.org is fine and has LE autorenewal. Add some setting to get the monitoring just on the host that currently has the tendril CNAME in DNS whichever it is, so that the role can be moved freely without causing false alerts.

https://icinga.wikimedia.org/cgi-bin/icinga/status.cgi?search_string=tendril ^ the green one is OK, but the yellow ones should be gone ``` dbmonitor1001 HTTPS-tendril This service problem has been acknowledged WARNING 2017-04-04 18:56:41 1d 17h 2m 52s 3/3 SSL WARNING - Certificate tendril.wikimedia.org valid until 2017-04-10 01:53:01 +0000 (expires in 5 days) dbmonitor2001 HTTPS-tendril This service problem has been acknowledged WARNING 2017-04-04 18:55:54 1d 17h 8m 40s 3/3 SSL WARNING - Certificate tendril.wikimedia.org valid until 2017-04-10 01:47:36 +0000 (expires in 5 days) einsteinium HTTPS-tendril OK 2017-04-04 18:56:42 54d 4h 48m 26s 1/3 SSL OK - Certificate tendril.wikimedia.org valid until 2017-05-28 03:56:00 +0000 (expires in 53 days) ``` This is because the monitoring check gets added per role::tendril, and moving tendril from einsteinium to dbmonitor is ongoing thing afaict. Multiple hosts get the check but only one actually has the tendril name in DNS, as of right now that is still einsteinium. So there is no real cert expiry issue here, the actual tendril.wm.org is fine and has LE autorenewal. Add some setting to get the monitoring just on the host that currently has the tendril CNAME in DNS whichever it is, so that the role can be moved freely without causing false alerts.