[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Feature request: UDP generic protocol testing

From: Alex Black
Subject: RE: Feature request: UDP generic protocol testing
Date: Tue, 25 Oct 2005 11:21:56 -0400

Hi Jan-Henrik, thanks for the response.

> Yes, we seems to need UDP read/write support, I can look into it.

That would be great, let me know if you need help testing.

> We have discussed this before and decided against a plugin  
> architecture. There are many reasons for this, one is that we want  
> monit to be autonomic.

Ok.  How about calling scripts, or executing programs? I guess that is
still a plugin architecture? You would need convention for calling out
to the script/program, and a way to communicate results back to monit.

I see that Monit does allow executing scripts/programs when a test fails
- so is it really autonomic? ;)

> > - As I emailed previously, being able to declare dependencies for  
> > monitoring of tests woudl be very helpful... Unless there is  
> > already a way to do this that I am missing :)
> This is a good idea, one way to achieve this is to scope tests in  
> blocks using {}, just like a standard programming language. 
> For example
>   if failed test {
>      if test
>   } else {
>      if failed test else if test
>   }
> What do you think?

That looks good, does this exist today in 4.6 ? Or are you proposing
using test scoping to make tests dependent on other tests?

Test scoping via { and } sounds decent. An alternate idea would be to do
something like you have for service dependencies, using the depends
keyword. Perhaps add a new keyword. I *thought* writing 'depends on foo'
would accomplish this.. I'm having trouble coming up with an alternate

Perhaps 'depends on test foo'?  Test would qualify depends to mean don't
test this unless the test foo passes.  

One other question. I want to be able to set a general alert email, 'set
alert foo1', and then on each test I would like to optionally be able to
override it, with 'alert foo2', but I've found in that cases emails get
sent to both foo1 and foo2. Is that the correct behaviour? Is there any
way to get the behaviour I am thinking off? (without doing separate
alerts for every test - cause having a per-test alert is not the normal
case for me)

> --
> Jan-Henrik Haukeland
> Mobil +47 97141255
> --
> To unsubscribe:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]