Archived community.zenoss.org | full text search
Skip navigation
2015 Views 1 Reply Latest reply: Apr 12, 2012 1:18 PM by jcurry RSS
Jim OBrien Newbie 1 posts since
Apr 12, 2012
Currently Being Moderated

Apr 12, 2012 12:32 PM

Small Consulting Opportunity

Hi All,

 

We are looking for some help with a Zenoss installation.  Below is a summary of what we have setup, where we are seeing, and in which areas we think we could use some help.  Also some questions for those people who might be interested and able to help us.

 

  We installed Zenoss ourselves, following the general documentation, made it work, have been happy with what it does for us.  We have about 75 devices that we are monitoring.  These are mostly linux blade servers and Cisco routers/switches. There are a few windows servers that we have added more recently, need to really understand their monitoring better.  Zenoss runs on it's own server, we also run RANCID on this server for doing config backups of the cisco devices, but generally it is on its own server.  The server is a 4 core, 4GB RAM, 10K speed disks. Feels like more than enough for Zenoss (perhaps?).  We are running Zenoss on RHEL5.5.  Most of the devices we monitor are local.  We have a very small set of remote devices that we monitor over a VPN.

 

  We've been using this setup for some time, and occasionally we get alarms that processes on servers have failed, or servers are down.   From looking at things what we can tell these are false alarms.  This isn't ideal, and is something we'd like to fix.  We don't really know our way around optimizing Zenoss.  We do often see zenmodeler take 100% of one CPU core.  Apparently this isn't multi-threaded or we're just not setup ideally?  This hasn't reached a pain point yet, but certainly we're at the point where I'd like to make things better.

 

  We would want someone to look through the config and make improvements. The areas would be: 1. fix what we haven't configured, or haven't configured properly with the installation / general overall setup. 2. Look at how we monitor blades / servers / processes and suggest improvements.  Generally we're using ping, SNMP, and process monitoring.  Perhaps in some places this is overkill.  3. Establish templates for new devices / clean up existing devices of each "kind". 4. bring us up to date as we haven't upgraded Zenoss in longer than I can remember. This feels not critical for me, but perhaps there are large performance improvements to be had.

 

  I have the impression that someone with a good working knowledge of Zenoss could likely find a handful of massively flawed things we are doing in a very short time and provide a huge benefit for the time spent.  This is mostly item #1 above.

 

  If you are interested / could help, can you tell me about your experience with Zenoss, experience / background in general, and experience with similar projects.  I would need to understand what the costs would be as that is a consideration.  How do you normally work?  Based upon above ( or please ask more questions via email ) what do you think the effort related to each area (items 1-4) would be?  Breaking it down into steps/stages may be important for us. 
  Please email me at the address below, obfuscated so I don't get too much spam.
Thanks,
Jim O'Brien
Director of Server Engineering
jobrien aT counterpath dOT com

 

  • jcurry ZenossMaster 1,021 posts since
    Apr 15, 2008
    Currently Being Moderated
    1. Apr 12, 2012 1:18 PM (in response to Jim OBrien)
    Re: Small Consulting Opportunity

    Hi Jim,

    I have a few clients for whom I have done something similar to what you are asking.

     

    You have probably found my CV around, but if not, go here -

    http://www.skills-1st.co.uk/ .  I have been around the systems

    management part of the industry for over 20 years, many of them around

    IBM and the last 5 years or so, I have focused much more on open source

    solutions, especially Zenoss.

     

    I do a mixture of teaching and consultancy and have developed a couple

    of Zenoss workshops that have run several times, both in classroom

    environments and using remote-teach technologies.  If you found me here on

    the Zenoss forum then you probably know that I have been contributing

    responses, documents and ZenPacks over several years.

     

    I have a number of clients for whom I have done similar Zenoss work to

    what you propose.  These days, I actually have more remote clients that I

    have never met face-to-face, than those that I actually visit!  We find

    that a combination of Skype and remote access to a client's Zenoss

    system, works well.  For access, we have a fixed external IP address; so

    some folk simply open ports 8080 and ssh for our source address to their

    zenoss server.  We have also used nx, VNC, Citrix and various VPN

    technologies.

     

    Typically an engagement starts with a healthcheck to understand your

    environment and provides written recommendations.  Typically this is 2

    days - depends on the size of the environment and how many issues there

    are with it.  Simple issues I will generally fix in the 2 days.

     

    Once we agree what needs changing, we can work in 3 ways:

    1) I fix it

    2) You fix it

    3) We fix it - generally I've used VNC as a screen-sharing technology

    though I have also used join.me

     

    Depends entirely on you.  My feeling is that folk ought to insist on

    skills transfer whenever they use a consultant, but often workloads

    don't permit.  Looks like you are US East Coast so the time difference

    shouldn't be a big issue.

     

    Some folk then setup an ad hoc support contract - typically between 1

    and 5 days - to fix problems as they arise.  We cannot offer a formal

    9-to-5 support contract as I have other clients and commitments but,

    other than planned holidays, I am generally contactable

    one-way-or-another, most days and will respond to issues as quickly as

    possible.  For such contracts, I generally provide a weekly timesheet

    for work done so we all know where we stand.  Billing is monthly or at

    the end of a contract, payable 28 days after invoice date.  Rates for

    remote work are £800 (GB pounds) / day.; "days" are typically made up of

    hours spread over a much longer period.  We will only bill for work

    actually done; if you setup a 5-day contract and you feel we have

    finished in 3 days, you get billed for 3 days.

     

    With regard to your points 1-4, I think the 2-day healthcheck will

    probably tackle points 1-3.  Updating Zenoss depends rather on where you

    are starting from and where you want to get to?  Also, the number of

    ZenPacks installed can have a big bearing on this.  If you are at a

    Zenoss 2.x then it is probably worth upgrading to 3.1 or 3.2.1 - depends

    very much on exactly what ZenPacks you have installed.  If you are at

    3.1, I personally would stay there.  3.2 wasn't good and although 3.2.1

    fixed many issues, it broke some more ZenPacks.  4 is currently in alpha

    but I wouldn't be anticipating putting that anywhere near production

    this side of summer.

     

    Does that help?

     

    Please feel free to discuss further.

     

    Kind regards,

    Jane