Discussion:
[Check_mk (english)] Central alert handling in distributed monitoring
Devin Kramer via checkmk-en
2018-11-08 02:33:29 UTC
Permalink
We currently have a distributed check_mk monitoring environment using the
"Raw Edition". Basically we have multiple check_mk nodes deployed in each
of our data centers each responsible for an equally distributed number of
hosts. We then have a single instance in our primary DC that is an
aggregated view of all of the other check_mk servers. The primary
motivations for this architecture was ensuring we could support the number
of nodes we needed to monitor in each location. Also, we like that
monitoring notifications are not dependent on a single system that has to
be reachable across possibly unreliable connectivity to a central DC. The
major draw back of this is we cannot leverage cross host or cross dc
dependencies because each check_mk nodes i not aware of the others.

At least at first glance of the product editions page
https://mathias-kettner.com/editions.html it would seem the Enterprise
edition has "central alert handling in distributed monitoring" and thus
maybe dependencies?

First, I am wondering if anyone has solved or mitigated this issue without
having to deploy the Enterprise Edition?

Second, I am wondering if anyone has experience with setting up
dependencies using the Enterprise Edition in an environment as large as
10K+ nodes 300K plus services across 10 locations?

Appreciate any insights folks are willing to share..

Devin Kramer
Andreas Döhler
2018-11-11 09:32:01 UTC
Permalink
Hi Devin,

The central alert handling in enterprise edition is what it says. Your
problem with cross dc dependencies can only be solved at the moment with BI
aggregations on your central system.
This BI aggregations can then be used as normal services and you can use
these services for notifications.

On my bigger installations i don't setup many dependencies. Most of these
things i do with BI.
At the moment i think there is no way to define real dependencies over more
then one instance, no matter if you use the RAW or Enterprise edition.
BI is working over many instances.

Best regards
Andreas


Am Do., 8. Nov. 2018 um 03:34 Uhr schrieb Devin Kramer via checkmk-en <
Post by Devin Kramer via checkmk-en
We currently have a distributed check_mk monitoring environment using the
"Raw Edition". Basically we have multiple check_mk nodes deployed in each
of our data centers each responsible for an equally distributed number of
hosts. We then have a single instance in our primary DC that is an
aggregated view of all of the other check_mk servers. The primary
motivations for this architecture was ensuring we could support the number
of nodes we needed to monitor in each location. Also, we like that
monitoring notifications are not dependent on a single system that has to
be reachable across possibly unreliable connectivity to a central DC. The
major draw back of this is we cannot leverage cross host or cross dc
dependencies because each check_mk nodes i not aware of the others.
At least at first glance of the product editions page
https://mathias-kettner.com/editions.html it would seem the Enterprise
edition has "central alert handling in distributed monitoring" and thus
maybe dependencies?
First, I am wondering if anyone has solved or mitigated this issue without
having to deploy the Enterprise Edition?
Second, I am wondering if anyone has experience with setting up
dependencies using the Enterprise Edition in an environment as large as
10K+ nodes 300K plus services across 10 locations?
Appreciate any insights folks are willing to share..
Devin Kramer
_______________________________________________
checkmk-en mailing list
Manage your subscription or unsubscribe
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
Devin Kramer via checkmk-en
2018-11-30 22:53:31 UTC
Permalink
Andreas,

Honestly we never even looked at the Business Intelligence features.
Going to need to do some reading it seems.. Any particularly good
documentation or guidance on the use of this feature you mind sharing.

Thanks
Devin
Post by Andreas Döhler
Hi Devin,
The central alert handling in enterprise edition is what it says. Your
problem with cross dc dependencies can only be solved at the moment with BI
aggregations on your central system.
This BI aggregations can then be used as normal services and you can use
these services for notifications.
On my bigger installations i don't setup many dependencies. Most of these
things i do with BI.
At the moment i think there is no way to define real dependencies over
more then one instance, no matter if you use the RAW or Enterprise edition.
BI is working over many instances.
Best regards
Andreas
Am Do., 8. Nov. 2018 um 03:34 Uhr schrieb Devin Kramer via checkmk-en <
Post by Devin Kramer via checkmk-en
We currently have a distributed check_mk monitoring environment using
the "Raw Edition". Basically we have multiple check_mk nodes deployed in
each of our data centers each responsible for an equally distributed number
of hosts. We then have a single instance in our primary DC that is an
aggregated view of all of the other check_mk servers. The primary
motivations for this architecture was ensuring we could support the number
of nodes we needed to monitor in each location. Also, we like that
monitoring notifications are not dependent on a single system that has to
be reachable across possibly unreliable connectivity to a central DC. The
major draw back of this is we cannot leverage cross host or cross dc
dependencies because each check_mk nodes i not aware of the others.
At least at first glance of the product editions page
https://mathias-kettner.com/editions.html it would seem the Enterprise
edition has "central alert handling in distributed monitoring" and thus
maybe dependencies?
First, I am wondering if anyone has solved or mitigated this issue
without having to deploy the Enterprise Edition?
Second, I am wondering if anyone has experience with setting up
dependencies using the Enterprise Edition in an environment as large as
10K+ nodes 300K plus services across 10 locations?
Appreciate any insights folks are willing to share..
Devin Kramer
_______________________________________________
checkmk-en mailing list
Manage your subscription or unsubscribe
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en
--
Devin Kramer Director, Infrastructure Engineering

12181 BLUFF CREEK DRIVE, 4TH FLOOR, PLAYA VISTA, CA 90094

RUBICONPROJECT.COM <http://www.rubiconproject.com> | JOIN US
<http://rubiconproject.com/join-us/> | FOLLOW US
<http://twitter.com/rubiconproject>
Continue reading on narkive:
Loading...