NETWAYS/check_vmware_nsxt

some ideas for improvement

Closed this issue · 10 comments

sklaes commented
  • Introduce a toggle to disable the proxy e.g. --noproxy
  • Alarms: Show performance data for high, critical, medium individual e.g. | alarms=12,high=5,critical=3,medium=4;;;0
  • Alarms: allow to include / exclude alarms e.g. --exclude=medium or --include 'Certificates/'
  • capacity-usage: allow to include / exclude checks, so that you can split it into multiple icinga checks
  • Alarms: the help message could point out, that only open alarms will be checks
  • Alarms: add a timefilter, so that you can create a volatile check in icinga, to get informed for every new alarm

and by the way, the releases page on exchange.icinga.com lists v0.1.0 as latest
https://exchange.icinga.com/netways/check_vmware_nsxt/releases

martialblog commented

Hi, thanks for the feedback. Most of that seems doable I think, but I have to double check.

Do you want to open a PR? Otherwise I cannot say when stuff will be implemented

sklaes commented

It was just a brainstorming on how to improve the script. I haven't implemented the check in Icinga yet and I'm not sure if/which of the above features I need, so I can't say if I will implement any of them

martialblog commented

No problem, feedback is always welcome.

  • 👍1
martialblog commented

@sklaes Could you elaborate what you mean with the "--noproxy" option?

sklaes commented

If there is an environment variable defined such as http_proxy, HTTP_PROXY, https_proxy or HTTPS_PROXY, python will use the proxy instead of a direct connection.
Having a lot of checks where some - but not all - need a proxy it is a pain in the a** to configure the no_proxy variable correctly. Something like this could make it a much easier:

if args.noproxy:
    os.environ['no_proxy'] = '*'
martialblog commented

I see. Woudn't unsetting the variable on specific nodes be an option?

I'm not sure yet if this is something that should be included in the code.

martialblog commented

@sklaes I'm not a user of the Plugin, can you give some details on this point?

add a timefilter, so that you can create a volatile check in icinga

sklaes commented

@sklaes I'm not a user of the Plugin, can you give some details on this point?

add a timefilter, so that you can create a volatile check in icinga

You could set your check interval and the time filter to e.g. 5 min, so you will only get new alerts. The check will automatically reset to green on the next check. (You also need to mark the check as volatile in Icinga to get a notification on the first state change)
This could be useful if you have alerts that you can't solve shortly, but setting a downtime or ack is not an option, because your do not want to miss any new alerts.
However on this check I don't think it is necessary to implement, because your can acknowledge alarms individually in NSX-T and thereby exclude it from the Icinga check.

martialblog commented

Thanks again. If need arises we could still do that, for now I'm not sure about implementing this.

martialblog commented

We implemented a couple of the suggestions in the new release. We'll update the Icinga Exchange soon.