1

Hello I been thinking about how to do a regular check on a system to assess the state of health of the systems. I am really turning into a system admin gradually and I already manage about 10 ubuntu servers. I have set up zabbix already but I believe either myself or a script should be run every month to just check on the servers.

Is it how it's done in real life

if yes what are the ways to achieve that?

is there any suggestion anyone feels sharing with a young system admin?

THanks for your advices

black sensei
  • 609
  • 3
  • 8
  • 25
  • Why run something "every month" when you already have something running continuously? – Michael Hampton Aug 16 '13 at 20:15
  • 1
    [Monitoring](http://serverfault.com/questions/44/what-tool-do-you-use-to-monitor-your-servers) is your friend – user9517 Aug 16 '13 at 20:18
  • 1
    What's wrong with your Zabbix install that you don't trust the data that it is giving you? – MDMarra Aug 16 '13 at 20:19
  • 2
    https://lopsa.org/mentor/WhyProtege – mfinni Aug 16 '13 at 20:27
  • 2
    You're already *Doing It Right* (you have Zabbix to be your monitoring system) -- my advice is don't start *Doing It Wrong* by adding periodic scripts, and consider *Doing It Better* by adding a ticketing system like [RT](http://bestpractical.com/rt) to receive your alerts and track how you resolve them. (Ticketing systems can also be used for other sources of work, like user requests -- they're really fantastic tools if set up and used properly!) – voretaq7 Aug 16 '13 at 20:37

1 Answers1

5

There is no need to add more clutter to your inbox by running a periodic script, and there is no need to create busywork for yourself by manually checking systems.

Zabbix is already doing regular checks. If it's properly configured it's doing them every 5-15 minutes (which is more frequently than you could ever do them manually), and it's alerting you when something needs attention (and ONLY when something needs attention).

You, as the sysadmin, should ensure that your automation (monitoring) works properly, and that it's monitoring what's important. If you want to be thorough, test it (fill a disk -- preferably a non critical one -- and be sure you get an alarm).

voretaq7
  • 79,345
  • 17
  • 128
  • 213
  • Great insight , thanks . I will look into the ticketing system.I think we have some license for kayako as well, am yet to look at what I can do for it.thanks – black sensei Aug 21 '13 at 07:27