Re: [Nagios-devel] NSCA patch submission - long plugin support,

[Guest]

On 29 Mar 2011, at 21:35, Mike Lindsey wrote:

> I've made a few modifications to NSCA, to allow both long plugin =
output support, as well as to bypass the command pipe.
>=20
> I've been using these patches in a production monitoring environment =
for over a year with about 1000 servers, 8000 service checks, with five =
minute retries.
>=20
> Performance tests for the long plugin output support (4k character =
limit, just like Nagios core) showed a very slight improvement in =
performance - I don't entirely know why, but I assume it has some arcane =
relationship with network packet sizes.
>=20
> Performance tests for bypassing the command pipe and submitting =
results directly into the check result directory have shown significant =
improvements in performance. I can supply some data if there's =
interest, but I suspect that some of the core developers have better =
environments for collecting and verifying that data.

We couldn't extend NSCA in a backwards compatible way so switched to =
using NRD instead. We got shockingly good performance too, so it's good =
to see changes to NSCA may also improve performance. Our results are =
here: =
http://labs.opsview.com/2011/01/next-ge ... oring-the=
-opsview-way/

Ton






This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]