1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Nagios Monitoring & Branches of?

Discussion in 'Servers and Enterprise Solutions' started by DHR, May 16, 2018 at 6:23 PM.

  1. DHR

    Mobster

    Joined: Apr 30, 2003

    Posts: 2,986

    Looking to overhaul my monitoring solutions and have had Nagios mentioned to me a few times.

    Anyone here use it in anger? Would be monitoring a mix of local Windows environments along with Azure & AWS infrastructure.
     
  2. BongoHunter

    Hitman

    Joined: Apr 14, 2014

    Posts: 733

    Nagios is widely deployed and so are its siblings like CheckMK and Icinga

    TBH its had its day and should be killed off as soon as possible

    If you want something modern I would be looking at Prometheus, with Grafana for frontend if you want dashboards

    Use the Prometheus Alert Manager addon to handle you alerts

    If you need logging analysis add ELK into the mix

    All are free, opensource, well maintained with good documentation.

    https://prometheus.io/
    https://grafana.com/
    https://www.elastic.co/
     
  3. DHR

    Mobster

    Joined: Apr 30, 2003

    Posts: 2,986

    Will have a look into those others thanks.

    Just seen auto correct decided to go nuts on the thread title too :-( if a mod is around can you change it to Nagios from nations! Thanks.
     
  4. Conanius

    Capodecina

    Joined: Oct 18, 2002

    Posts: 12,164

    We've recently moved from Solarwinds to a homebrew combination of Dashing, Icinga and Grafana.

    Superb set and wish we'd done it forever ago.

    That said, don't underestimate the effort to stay on top of it all.
     
  5. DHR

    Mobster

    Joined: Apr 30, 2003

    Posts: 2,986

    That's what I'm most concerned about. I need something that is largely set and forget, was waiting for status cake to get their windows monitoring sorted but they've let us down.
     
  6. BongoHunter

    Hitman

    Joined: Apr 14, 2014

    Posts: 733

    You need to automate everything you can with monitoring - manually rolling out agents, check's and alerting configuration is not supportable.

    So really you want agents/instrumentation rolled into your build process, and that same process also making any required updates to your monitoring config to pickup new endpoints to scrape etc.

    E.G if you use Kubernetes or say OpenStack, get a script to update a configuration file for prometheus in Git or some other SCM tool at the moment of deployment, and have something like Jenkins watching the repo and doing the deploy of configuration file to the prometheus box for you.

    Similar stuff can be done for Nagios with its hosts.cfg and other associated files, or with Icinga you have your build process call the rest API to add the host into monitoring.

    Obviously will all depend hugely on the size of your environment, the technologies used and probably most importantly the culture and expertise available in your organisation. It takes a bit of re-education and shift in mindset for people from a traditional ITIL/ITSM type support model moving into this new style of agile/devops/modern/new-style stuff (delete as appropriate!)
     
  7. BongoHunter

    Hitman

    Joined: Apr 14, 2014

    Posts: 733

    If your a wintel shop - do not be drawn to SCOM, I'm sure some organisations have made it work, but most do not get the anticipated level of success, and it's a ****** to support and move away from
     
  8. DHR

    Mobster

    Joined: Apr 30, 2003

    Posts: 2,986

    SCOM ..... possibly the worst monitoring product for SMBs :)
     
  9. DHR

    Mobster

    Joined: Apr 30, 2003

    Posts: 2,986

    I get that, it feels like there is no middle ground at the moment. I suppose it's a case of where to start right now. You have an existing estate, you need to monitor it, surely with the discussed solutions you can script up agent / config installation without integrating it at the start of the build process?
     
  10. BongoHunter

    Hitman

    Joined: Apr 14, 2014

    Posts: 733

    It's tricky for existing stuff if you have a mix of kit and configs, if youve got any deployment tools already for pushing stuff out to your estate then use it, if not then a bit of powershell to pull down agents and the bits you need to each box is probably the way to go, and template the agent configs by server role and OS version
     
  11. Caged

    Capodecina

    Joined: Oct 18, 2002

    Posts: 22,463

    If you want something easy and aren't fussed about it being free then PRTG is really good as well