HN Explorer

Ask HN: What are the monitoring tools you use?

2 points by noobplusplus ↗ HN
I am a newbie, and wanted to add monitoring services. I read about nagios, then later read more and found sensu|sonian was better at it http://www.sonian.com/cloud-monitoring-sensu/

Then there are things like new relic(which appears too complex to me)

I am a new sysadmin, and do not have much experience, though I have googled and read stuff, and done the homework.

I just wanted to ask folks over here, as to what services they use for monitoring their different servers, and what led them to their respective choices?

Thanks for reading the question.

6 comments

[ 3.2 ms ] story [ 22.9 ms ] thread
txutxu
Nagios is great if you manage bare metal.

If you have a very dynamic environment (a.k.a. virtualisation, cloud, etc) usually you have to go one level furter and integrate your provisioning tools with the monitoring (and backups) system.

I've never used sensu, but I've write my own monitoring and reporting tools many times. I can recommend you to try to get that skill, learn howto script a protocol and report, etc... probably you will need it to adapt your monitoring tool, whatever you choose.

txutxu
Also, as you ask for experiences:

In my current and previous job, nagios, with nrpe.

Before that, in a shop I did have only Debian servers, I did write my own dashboard.

The reason has been always pragmatism.

otisg
We use:

* SPM for monitoring and alerting for our ElasticSearch, Solr, Hadoop, and HBase clusters, as well as Java (web)apps + the usual server metrics + custom business and performance metrics

* Sensu for server metrics monitoring + alerting

URLs:

* http://sematext.com/spm

* https://github.com/sensu

troels
New Relic is primarily a tool for monitoring and diagnosing on the application lauer, although they have recently expanded their offerings.

For monitoring servers, I'm currently using munin and pingdom.

WestCoastJustin
I guess I need to know more about what you want to monitor. We use nagios (host/service up/down notification), ganglia (host metrics like ram/disk/etc), and MRTG for bandwidth etc, and rsyslog with tenshi for log alerting. If you are a sysadmin then you probably want nagios. I would be suspect of new shiny things, nagios is rock solid and is used as an industry standard for monitoring in many open source shops (think universities, HPC clusers, etc).

p.s. I run a sysadmin screencast site @ sysadmincasts.com. I'll try and do a nagios episode in the next couple weeks.