MySQL Bugs: #45981: Detail Error Messages in Dashboard for Agent/Server down

Bug #45981	Detail Error Messages in Dashboard for Agent/Server down
Submitted:	6 Jul 2009 20:20	Modified:	17 Dec 2014 17:15
Reporter:	Donna Harmon	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Enterprise Monitor: Web	Severity:	S4 (Feature request)
Version:	2.1.0	OS:	Any
Assigned to:	Assigned Account	CPU Architecture:	Any

Description:
Rather than just mark the agent as down (as you do now), or the db as unavailable (crossed out) you should provide a mechanism for the agent to report the ACTUAL problem to the dashboard.  That way there is only a single place to have to look to find the problem which makes it likely to recognize and be able to fix the problem earlier.

How to repeat:
When viewing the dashboard there is no indication as to why the agent and/or server is down/unavailable.

Hi. Just to be clear I see various issues:

1. The agent isn't reporting in normally so the dashboard/front-end does not know why. That's a valid situation (the sysadmin/dba forgets to start the agent)
2. The agent connects but is told to "go away" (duplicate agent/server). This can be reported on the dashboard and the information is useful
3. The agent is running "normally" but unable to connect to the database:
   3.1. it can not reach the db (db is down)
   3.2. it can not login to the db (wrong credentials)
   3.3. another reason ?

These different situations IF reported on the dashboard make diagnosis much clearer.

Maybe something like 'agent down, 1 mysqld pid found'? This doesn't guarantee that the pid belongs to the monitored instance, but it might help. 

Although it could link the pid and the instance by comparing the path in the commandline options.

This was resolved with the 3.0.0 release, via both the MySQL Availability Advisor (which reports Unavailable due to some connection error, or Down for no response), and the new Bad Connections panel within the MySQL Instances dashboard.