perfSONAR User Q&A and Other Discussion

[Danial Ebling <>]

We are running a 5.0 beta/test deployment with this strategy, where we have an additional archiver configured with our measurements. We’re using the RabbitMQ archiver plugin to connect to a Telegraf instance, where it then formats the measurement results and sends it our InfluxDB database where our other performance metrics are kept. It seems to work well, and the plugin is pretty configurable with timeout/retry settings. There’s also a Kafka plugin if that message bus is preferred.

 

Danial Ebling

Utah Education Network

 

From: <> on behalf of Mark Feit <>
Date: Thursday, December 15, 2022 at 1:09 PM
To: Jarett DeAngelis <>
Cc: <>
Subject: Re: [perfsonar-user] perfSONAR and Grafana

Jarett DeAngelis writes:

Actually, quick question: the docs say, for example:

…

pScheduler can be configured to apply an archive specification to every run it performs on a host  …

Okay, so on which grid host would I do this? I thought since our central management server (which I understand is also going away as we understand it currently in 5.0) currently collects all this data it might be on there, but nope.

 

My advice is don’t.  Global archiving was designed for a specific use case where the system(s) were part of multiple meshes over which the user had no control and they wanted to siphon off all of a particular type of measurement.  If you were going to use that, you’d have to install a file on every system that made your measurements.

 

If you’re using a pSConfig-based mesh, your best bet is to add whatever additional archiving you want to do to each measurement alongside sending it to Esmond.  You can archive to as many destinations as you want.

 

HTH.

 

--Mark