4
votes

In my office I have all our monitoring tools setup and working great, but I dont have an easy way to view them.

I have a large TV in my office plugged into an old PC that has my nagios status page always showing. If I need to change anything on that computer I use Synergy to access the computer from my desktop.

We are thinking about adding another TV and I want some suggestions on setting it all up.

We are a 99.99% Windows shop.

What do you use to run all the TV's in your helpdesk?

Synergy works for me, but what if one of the other admins want to change the screen? Is there any easy way that any of us (currently 4 people) can change the screen from our desktops? (Remote desktop doesn't work because it locks the console which is the output to the TVs.)

Any advice would help, Thanks.

8 Answers 8

4
votes

You could use VNC. It can be configured to not lock the screen and to allow multiple simultaneous users.

2
votes

Simple, just get organised. From the sounds of it you must be spending most of your time looking at screens, instead of using the computers to tell you when something need to be looked at.

Start by learning what really needs to be monitored constantly. e.g. There should be no reason to keep a Nagios display on screen if you set up alerting properly. Then reduce the clutter and learn to work with windows, rather than screens. Why is Nagios on a separate screen instead of a separate tab in your web browser? I'm sure the same principles can be applied to much of what you use.

3
  • 1
    I mostly agree with John about Nagios alerting. What the screen is really for is for the people manning the phones to look at. The people really fixing the problem already knew because the alarms went off, but every help desk tech doesn't need that information. The 1 of 4 Nagios screens lets them look for the 50 flashing red boxes that say "Leave the server people alone because they are fixing the problem." rather than having them ask. Dec 29, 2010 at 23:39
  • @Keith, the normal Nagios screen is not the best way to see that information. What you need is a system to list what alarms have gone off and what is currently being worked on. e.g. A system which grabs that information and sends it to a chat program would allow the helpdesk staff see a live running display in whatever chat client you decide to use to display it, which puts the relevant information where it needs to be - right in front of the operator. Dec 30, 2010 at 1:00
  • 1
    Valid. We don't display the normal screen, only the problem screen, which is normally pretty short and has few or no "reds". Mainly we use it for showing level 1 that there is a widespread problem and they can further research specifically which services are affected through other methods if it's important to their job. To be blunt about "what is currently being worked on", if it's broken, the server staff is working on it. If it's not working, we're not making money so we handle accordingly. Not the answer for everyone, but Blake did ask for "how do you do it?" Dec 30, 2010 at 3:59
1
vote

We have found allowing multiple people to control what is viewed hasn't worked so well. Some will want to look at their own task list, some will watch movies, right-handed (or -brained) people will want certain items on one side and others the opposite. As long as someone can rearrange it, they will and just as soon as you need to look at something when the stuff hits the fan it either won't be on the screen or in the wrong place.

We have 1 monitor watching the servers and 3 monitors split around the room watching the call and workflow queues.

1
vote

We use hobbit as our monitoring tool, its non green page has almost everything we need. But its lacking a view of.our vmware enviroment and everyting regarding ou inhouse.software. So, I'm probably going to write something from scratch.

I also recomend googling atlassian wallboard contest, some amazing entries there.

1
vote

The standard Nagios screens are not very handy for this kind of operation.

Using something like Nagvis to create your own view of the system may save you a lot of time, money, effort, wallspace and electricity. For various reasons I ended writing my own front-end - it's not hard!

Configure your email alerts.

Add an RSS feed.

Bombarding people with data via a single channel is counter-productive. Provide people with information and a choice about how they access it and they will be more involved, motivated and effective.

1
vote

We created a world map in Nagvis which shows status in each of the regions we operate in. Clicking on a region status icon brings up a larger scale map of the region, showing the status of key devices in each office.

We also created schematics showing dependencies of our central systems, and overlayed status icons for each part of the system.

Neither of these are hugely useful to my team since we get alerts telling us what is going on, but it is very useful to management to get an overview (they get excited when everything is showing as green - hell, so do I :-) Our support desk can also take a look at the various Nagvis pages instead of asking us "is there a problem with X?"

We did try displaying this on a screen which everyone could see, but it wasn't that useful. Either we need a bigger screen, or a way to drive people to look at the appropriate Nagvis pages from their own workstation.

1
vote

You don't. Half the point of nagios is to configure when you care and to have it monitor things for you. Typically by email. If you get too much email, I suggest configuring your client to treat CRITICAL status mail as special enough to pop up a window (unlike the rest of your incoming mail).

If you must have a status page, make it summarize the status. From my last job, I recall Opsview had a couple of nice pages, and the web frontend has parameters to where you can remove OK messages and only see the warning / unknown / critical services and hosts.

0
votes

In our office we use monit. Monit has a web interface that exposes the stats of the server as xml and json.

We poll this data every N minutes with a simple ruby script and create a webpage that has a listing of all the processes. If a service has a problem, then it is highlighted in red.

In theory, that page should have no service highlighted in red.

You can fit tens or hundreds of processes in one page without problem. You only act if you see glance to the monitor and see something in red.

Pretty effective for us.

1
  • Sorry, just read the 99.9% Winodows part.
    – edmz
    Jan 5, 2011 at 22:42

Not the answer you're looking for? Browse other questions tagged .