Re: Hosts on M/Monit keep going offline and online at regular intervals

[Abdul Munim Kazia]

I just checked a few of my servers which have had this issue. 

We are using Monit 5.5 and m/monit 2.4. 

The monit poll time is 15 seconds. 

I have changed the Acceptable Report Skew from 3 to 10 for these servers. I will monitor it for a day to see if the issue persists for these servers. I will probably send the logs across if this doesn't fix it. 

Also, I ran monit -v on one of the hosts, and I saw the following:

 M/Monit(s)         = http://192.168.48.1:8080/collector with timeout 5 seconds using credentials

Where is this 5 second timeout set? Will increasing this timeout fix the issue?

Abdul Munim Kazia
address@hidden

On 23 January 2014 15:53, Martin Pala <address@hidden> wrote:

I think this isn't scalability issue - we test with far higher load then 100 hosts (although in the long term we may add load balancing support).

Regards,

Martin

On 23 Jan 2014, at 11:04, Roose, Marco <address@hidden> wrote:

Hi,

I think this is a really interesting topic. One should think about the implementation of load balancing (and with that High Availability) into m/monit. Is there anything focusing into that in the development queue?

 

Best regards,

Marco Roose

 

From: address@hidden [mailto:address@hidden] On Behalf Of Abdul Munim Kazia
Sent: Thursday, January 23, 2014 10:38 AM
To: address@hidden
Subject: Hosts on M/Monit keep going offline and online at regular intervals

 

Hello

 

I have been using monit and m/monit at my organization for the past year or so. What started out as a simple setup with a handful of servers has now become a full monitoring dashboard with more than 100 hosts listed on m/monit. 

 

While monit itself works perfectly, at every regular interval of a few minutes, some of the hosts get reported as "no report from monit" in the m/monit console. The report comes in a few seconds and the host becomes green again. This happens fairly regularly, and it happens to a different set of hosts every time. All the hosts and the m/monit host itself are hosted in the same data center, so I don't think that network latency is an issue. 

 

This could be occurring because m/monit doesn't receive the data for all hosts in time, and it wouldn't be an issue for me, if it didn't crowd the events list with 20-30 events every half an hour. 

 

I have looked at the server.conf file, but I don't think any of those configuration settings will help me out, if I am not wrong. Has anyone faced this issue before? Is there any way to fix this, either from monit's or m/monit's end? Can I increase this timeout?

 

Thanks for you help

 

Abdul Munim Kazia

munimkazia.com

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general