perfSONAR User Q&A and Other Discussion

[Danial Ebling <>]

Thanks Mark, that is pretty much how our infrastructure is set up. In our case, we’re sending the perfSONAR measurements to an InfluxDB database because that’s where we keep our router telemetry data. Then Grafana (and other tools) query the data with the InfluxDB datasource plugin.

 

Danial

 

From: Mark Feit <>
Date: Tuesday, January 3, 2023 at 10:16 AM
To: Jarett DeAngelis <>, Danial Ebling <>
Cc: <>
Subject: Re: [perfsonar-user] perfSONAR and Grafana

Jarett DeAngelis writes:

 

can you say a bit more about how you set this up? I think I may be missing a middle step here. I thought I could use the Infinity data source plugin in Grafana, but I think that's not right, it's not able to listen to things posting to a URL.

 

I’m not familiar with the Infinity plugin, but a quick look at the docs hints that it’s really a many-common-formats-to-Grafana translator.   For using Infinity, the missing middle step is having a store for the data to live that can be queried.

 

RMQ and Kafaka are useful if you already have infrastructure in place for getting the information to whatever systems are going to hold and query it.  For example, WLCG has its own storage and display systems because they gather more than just perfSONAR data.  RMQ is its entry point.  They used to have a program that would query Esmond for new perfSONAR measurements and dump them into RMQ.  Now they have archivers in their meshes that archive directly to a RMQ topic that understands pScheduler’s test results.   (I’m not sure they’re fully migrated, but that’s where they’re headed.)

 

If you’re just starting out, I’d recommend posting to Elastic via HTTP and, optionally, LogStash.  We’ll be shipping something similar with 5.0.

 

--Mark