Bug #45937 server 'down' reported from mismatching UTC date/time of server and agent
Submitted: 3 Jul 2009 12:44 Modified: 21 Jul 2009 14:34
Reporter: Carsten Segieth Email Updates:
Status: Closed Impact on me:
Category:MySQL Enterprise Monitor: Server Severity:S3 (Non-critical)
Version: OS:Any
Assigned to: Eric Herman CPU Architecture:Any

[3 Jul 2009 12:44] Carsten Segieth
If agent's and server's time settings are out of sync the dashboard reports server "down".

How to repeat:
have a date/time settings of agent and server box with a 'big' difference, e.g. 1 h or so.

Suggested fix:
Server should check the UTC times sent in by the agent and report / log differences that are out of a 'resonable' difference.
[3 Jul 2009 14:36] Eric Herman
Now track a "Reporting Delay (ms)": the difference (in milliseconds) between the MEM Dashboard's time when it receives the agent request, and the data in the "utc" field of the agent's request packet.

Added the "Reporting Delay (ms)" to the UI both on the main dashboard page with a single server highlighted, as well as the Manage Servers page.

Logs a warning once (per agent restart) if more than 1/2 hour out-of-synch.
[6 Jul 2009 21:45] Enterprise Tools JIRA Robot
Keith Russell writes: 
Patch applied in versions =>
[9 Jul 2009 21:11] Enterprise Tools JIRA Robot
Carsten Segieth writes: 
tested OK in, the log message shows the difference and also the Meta Info now contains this info.
[21 Jul 2009 14:34] Tony Bedford
An entry was added to the 2.1.0 changelog:

If the Agent and monitored server had time settings that were out of sync the Enterprise Dashboard reported the server as “down”.