TECHZEN Zenoss User Community ARCHIVE  

zenoss AWS amazon ec2

Subject: zenoss AWS amazon ec2
Author: [Not Specified]
Posted: 2016-03-30 03:53

Hi there,

We were thinking of migrating our zenoss to latest version as an ec2 instance.

We only monitor 50 or so snmp routers. I find the requirements to be very steep to run zenoss...

20G of RAM, 500G of storage in all (SSD as DOC says) . Quad CPU , this Amazon ec2 instance would cost us 700$ or more a month !

Is there a way to run with lesser specs BTW, it's a question coming from my management...

thank you



Subject: > would cost us 700$ or more
Author: Jan Garaj
Posted: 2016-03-30 08:35

> would cost us 700$ or more a

Really

AWS:
r3.xlarge (4CPU, 30GB RAM) - on demand = ~240$/month
500GB SSD EBS = ~50$/month

Is it still expensive
- use AWS reserved instance, not on demand instance (~100$/month) - check AWS simple calculator
- start with small EBS volume - it's elastic volume so you can increase size anytime
- use Google Cloud not AWS (BTW Google has better networking and it can be also cheaper)
- use own HW
- maybe ZaaS will be cheaper - dunno https://www.zenoss.com/products/on-premises-vs-cloud
- use free resources of your high spec instance for (Bitcoin) mining - you can earn also some money :-D
- start with high spec and configure all devices in the Zenoss and then check memory usage - maybe you will be able to use also instance with 15GB memory (keep in mind that memory can be main bottle neck); if you accept also some risk, then you can probably also reduce number of HBase region instances https://support.zenoss.com/hc/en-us/articles/205319305-How-to-Reduce-the-Number-of-HBase-RegionServers-in-Zenoss-Resource-Manager

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

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



Subject: I don't know. Use classic
Author: Jan Garaj
Posted: 2016-03-31 10:53

I don't know. Use classic engineering process: try it and test it. Start 500GB EBS install what do you need, start to collect all metrics and measure disk usage. Then you will see what do you really need. Test also resizing before going to production.

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

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



Subject: Do you really want to go to
Author: Jane Curry
Posted: 2016-04-07 09:14

Do you really want to go to Zenoss 5 for what you are monitoring You may find far less pain currently sticking with Zenoss 4.2.5 - SUP671 is the latest patch level. Very stable.

Cheers,

Jane

Email: jane.curry@skills-1st.co.uk    Web: https://www.skills-1st.co.uk



Subject: Bennethos,
Author: [Not Specified]
Posted: 2016-04-21 02:25

Bennethos,

I'm going to agree with Jane. You'll have a less painful, better supported, most customizable and smaller foot print solution using Zenoss 4.2.5 for about 50 network devices.

Tehncially it's possible to run Zenoss 5 on something as limited as a Intel Nuc whichcosts a few hundred dollars to buy. It's just very tight. I'd very much reccomend 4.2.5 for this case.

ZenMaster Shane William Scott (Hackman238)

CTO

GoVanguard Inc.

sscott@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



< Previous
How complex is it to add buttons (or menu items) to the Event console?
  Next
Zenoss 5.1.1 - Trigger through Command action
>