Re-think availability processing

Heiko W.Rupp hrupp at redhat.com
Tue Jun 1 12:25:28 UTC 2010


While testing some stuff, I had my metric collection set to 1min and at one point had
to shut down the agent.

Result was that the server correctly saw the agent going down. Metric did not show up equally
as expected. BUT just showed as green. 
Of course one can argue that the monitored resource itself was not down, so green is correct.
But then the graph just having a hole in the metrics but always showing green is confusing
(se attachment)

Another area where our users are struggling and which lessens the value is the relatively long
availability reporting interval, which we've introduced to reduce the load for large scale systems.
I know that this can be configured; still operators expect to hear from downed hosts earlier from their
SLM system than from the customers.

The third area (which is my pet) is the issue of lingering resources that are always down, just
because they got auto detected, but are not in use and can't be auto-inventoried.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: Bildschirmfoto 2010-06-01 um 12.52.42.png
Type: image/png
Size: 23548 bytes
Desc: not available
Url : https://fedorahosted.org/pipermail/rhq-devel/attachments/20100601/bdb6a546/attachment-0001.png 
-------------- next part --------------

-- 
Reg. Adresse: Red Hat GmbH, Otto-Hahn-Strasse 20, 85609 Dornach bei M?nchen
Handelsregister: Amtsgericht M?nchen HRB 153243
Geschaeftsf?hrer: Brendan Lane, Charlie Peters, Michael Cunningham, Charles Cachera



More information about the rhq-devel mailing list