Thanks Heiko for your response.  I guess, your solution should address my problem correctly. However, does RHQ support data trending purely based on the metrics that it collects ?

For eg: Lets say we have a JMX attribute that reports the live transaction count, incremented everytime there's a transaction happen in a system. Can I use RHQ's collection and plotting techniques to identify peak traffic hour during a day ?

I believe, RHQ is good at plotting the metrics that we expose through our application but for trending data we should feed the correct data( that is actual trending data ) to it from within the application. For ex: Active thread count in Jboss.

Am I missing something here ?

Thanks ,
Sarat kumar.

On Thu, Jul 14, 2011 at 5:30 PM, <rhq-users-request@lists.fedorahosted.org> wrote:
Send rhq-users mailing list submissions to
       rhq-users@lists.fedorahosted.org

To subscribe or unsubscribe via the World Wide Web, visit
       https://fedorahosted.org/mailman/listinfo/rhq-users
or, via email, send a message with subject or body 'help' to
       rhq-users-request@lists.fedorahosted.org

You can reach the person managing the list at
       rhq-users-owner@lists.fedorahosted.org

When replying, please edit your Subject line so it is more specific
than "Re: Contents of rhq-users digest..."

Today's Topics:

  1. Need suggestions to implement a trend-like metrics (Sarat Beesa)
  2. Re: Need suggestions to implement a trend-like metrics
     (Heiko W.Rupp)


---------- Forwarded message ----------
From: Sarat Beesa <sarat.beesa@gmail.com>
To: rhq-users@lists.fedorahosted.org
Date: Wed, 13 Jul 2011 19:05:29 +0530
Subject: Need suggestions to implement a trend-like metrics
All,

I've been using RHQ for past couple of days to integrate our application. For a newbie like me, RHQ was very easy to learn and ponder around ( UI is very intuitive ). I've been able to get it working very quickly with my application. Thanks to you guys for coming up with such a great application.

Now, I have a situation, where I should demonstrate that RHQ should raise an alert in case 'If connectivity to an external interface fails for 3 times in past 5 mins" raise a medium priority alert to a given list of Users. To demonstrate this I have instrumented my code and exposed a JMX attribute to denote this value.  The value of the JMX attribute is incremented everytime we have a connectivity timeout with the external interface.

RHQ was able to fire an alert, when this counter is reaching the value of 3 but since this value is not reset from my application it was never able to recover from this situation. For ex: After 3 connection timeouts, the connectivity to the external interface resumed, so the counter stays at 3 even for the next collection of metrics, next time RHQ queries the metrics it identifies the value again matches the rule and it fires an alert. What is the best way to handle in this kind of situation ?

Ideally, if RHQ can measure the difference between two successive collections, then it can plot a trend and identify the behavior that connectivity has been resumed after sometime. Can RHQ support collection of metrics in this way ?

Any suggestions are welcome.

Thanks,
Sarat kumar.


---------- Forwarded message ----------
From: "Heiko W.Rupp" <hrupp@redhat.com>
To: rhq-users@lists.fedorahosted.org
Date: Wed, 13 Jul 2011 16:37:05 +0200
Subject: Re: Need suggestions to implement a trend-like metrics
Sarat,

> I've been using RHQ for past couple of days to integrate our application. For a newbie like me, RHQ was very easy to learn and ponder around ( UI is very intuitive ). I've been able to get it working very quickly with my application. Thanks to you guys for coming up with such a great application.

Thanks

> Now, I have a situation, where I should demonstrate that RHQ should raise an alert in case 'If connectivity to an external interface fails for 3 times in past 5 mins" raise a medium priority alert to a given list of Users. To demonstrate this I have instrumented my code and exposed a JMX attribute to denote this value.  The value of the JMX attribute is incremented everytime we have a connectivity timeout with the external interface.
>
> RHQ was able to fire an alert, when this counter is reaching the value of 3 but since this value is not reset from my application it was never able to recover from this situation. For ex: After 3 connection timeouts, the connectivity to the external interface resumed, so the counter stays at 3 even for the next collection of metrics, next time RHQ queries the metrics it identifies the value again matches the rule and it fires an alert. What is the best way to handle in this kind of situation ?

Just send a 1 when the connection is not available and a 0 else.
Then define an alert that triggers on value > 0.5 and define a dampening rule for N occurrences in X minutes

Hope that helps
 Heiko

--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschaeftsführer: Brendan Lane, Charlie Peters, Michael Cunningham, Charles Cachera



_______________________________________________
rhq-users mailing list
rhq-users@lists.fedorahosted.org
https://fedorahosted.org/mailman/listinfo/rhq-users