Zenoss ZenTech Community

I am quite new to ZenOSS Core and I am a bit confused about the ZenOSS Core vs Commercial comparison on zenoss.com
Is there a hard limitation to 1000 Devices in ZenOSS Core

Subject:	Hi!
Author:	Andrew Kirch
Posted:	2014-11-03 08:02

Hi!

Zenoss Core is open source. There is no hard device limit. The reason why 1000 devices is considered a "cut-off" is non-technical, but represents a point where the labor cost of maintaining Zenoss makes a supported commercial version set up by our folks desirable.

Andrew Kirch

akirch@gvit.com

Need Zenoss support, consulting or custom development Look no further. Email or PM me!

Ready for Distributed Topology (collectors) for Zenoss 5 Coming May 1st from GoVanguard

Subject:	No hard limit, just
Author:	[Not Specified]
Posted:	2014-11-03 08:02

No hard limit, just limitations with a single server hosting all devices. You can install additional collectors, hubs to accommodate >1000 devices. I personally am running 1600 devices ( all routers/switches) on a single core install. What the commercial install provides you is support for large architectures and enterprise type zenpacks (althought some of these are now available for core).
--Rob

Subject:	There's at least 4 people I
Author:	Andrew Kirch
Posted:	2014-11-03 08:03

There's at least 4 people I know of in the community monitoring more than 1000 devices on Core, so it does work.

Subject:	1755 here, using 1 Zenoss
Author:	[Not Specified]
Posted:	2014-11-03 10:07

1755 here, using 1 Zenoss master and 6 distributed collectors.

Subject:	1755 is a lot, way to go!
Author:	Andrew Kirch
Posted:	2014-11-03 14:06

1755 is a lot, way to go!

Subject:	Thank you guys, now it's
Author:	[Not Specified]
Posted:	2014-11-26 16:01

Thank you guys, now it's clear! :)

Subject:	Ah, just a further question:
Author:	[Not Specified]
Posted:	2014-11-27 03:16

Ah, just a further question:

Do you use ZenOSS also for log management in these large environments

BR & Thanks!
lousek

Subject:	Hi again, good question again
Author:	Andrew Kirch
Posted:	2014-12-01 09:14

Hi again, good question again!

Most larger installations will filter logs through something like logstash before processing them with Zenoss to avoid having Zenoss churn through every log entry.

Subject:	2961 Devices here.
Author:	Brad
Posted:	2014-12-02 14:30

2961 Devices here.

Not doing any log management. But am a large amount of custom monitoring. I am monitoring very large switches, routers, Linux, Windows, Solaris, etc. with many OID's. Running 4 pollers and a single instance as the HUB with no issues. The trick is knowing how to tune it and how to setup the pollers. I am shooting for 4,000 devices before I get a little ansy to move to remote hubs.

As for the cost savings as so stated I would say it's more around the 10,000 device range. Once it's up it just runs. It's not like they are going to be cleaning up old devices and adding new ones for you. We actually moved from Enterprise back to Core because the enterprise zenpacks weren't even being used and the support we were asking for wasn't worth the cost. They need a better pricing model.

Subject:	2600 devices and counting
Author:	[Not Specified]
Posted:	2014-12-04 17:12

2600 devices and counting

We are not doing any log management with Zenoss either. We are also looking into Enterprise support as well. We've recently migrated to 4.2.5 and have been finding ZenPacks for Enterprise that we are interested in. I personally have not been in contact with Zenoss about support but from what I hear the pricing model is not the best.

Serpent, I would be interested in your server configuration. We're getting to the point where the hard drives in our box can't keep up with all the rrd files that need to be written and have seen 100% load on the hard drives at times.

Subject:	Just a quick FYI we've
Author:	Andrew Kirch
Posted:	2014-12-08 16:11

Just a quick FYI we've revised our pricing model in the past few months, it might benefit you to reach out and ask!

Subject:	Baileytj,
Author:	Brad
Posted:	2014-12-10 16:38

Baileytj,

One word, ramdisk. Our servers have 48GB+ in them. I have set aside 16GB (only using 7GB or so) with a ramdisk. Then I sync that ramdisk every 30 minutes to the local drives. Takes like 30 seconds is all. Then I have a script setup to copy from the local disk on bootup. It increases performance tremendously.

Don't get me wrong, I understand needing support now and then. But at least offer more than just some phone call's and a few little fixes here and there. For the price I would of expected them to write custom zenpacks and to help with some other things. That would of made more sense and they could of had it so they could offer those zenpacks to others. Just my 2 cents.

Additionally. The monitoring calculator for support costs looks to be made up by marketing. For this setup you don't need a dedicated DBA, SME, Business Analyst. This isn't IBM Netcool or the like.

Subject:	Serpent, we've been thinking
Author:	[Not Specified]
Posted:	2014-12-10 16:50

Serpent, we've been thinking about that. Glad to hear it's been done before.

Subject:	Serpent,
Author:	Jan Garaj
Posted:	2014-12-12 04:31

Serpent,

What is advantage of RAM disk I'll be happy if you can provide real metrics (HDD queue/utilization/svctm/..., Zenoss response times, ...).

IMHO:
1.) if you have enough RAM, then free RAM is used as filesystem cache, so Zenoss python files will be cached in the RAM anyway, my case:
root@zenoss:~# free -g
total used free shared buffers cached
Mem: 47 44 2 0 3 21
-/+ buffers/cache: 19 28
Swap: 0 0 0
=> 21GB of RAM is used for file caching
2.) if you use rrdached, then rrdcached caches RRD IOps and is writing RRD to disk every 30 minutes (depends on the settings), so again IOps are cached in the RAM
=> you are wasting your RAM :-)

I'm solving problem with IOps peak, because I have high disk queue (~0.5k), disk utilization (~70), await time (0.5s) during rrdcached writing. I'm thinking how to solve it:
- SW tuning of block devices: settings nr_requests and queue_depth
- HW tuning -> SSD
IMHO: RAM disk doesn't help

Devops Monitoring Expert advice: Dockerize/automate/monitor all the things.

DevOps stack: Docker / Kubernetes / Mesos / Zabbix / Zenoss / Grafana / Puppet / Ansible / Vagrant / Terraform / Elasticsearch

Subject:	ZenOSS Core - Device Limit
Author:	[Not Specified]
Posted:	2014-11-01 08:27

ZenOSS Core - Device Limit