Skip to Content.

rare-dev - Re: [rare-dev] Prometheus DB and Grafana Dashboards

Subject: Rare project developers

List archive


Re: [rare-dev] Prometheus DB and Grafana Dashboards


Chronological Thread 
  • From: mc36 <>
  • To: Tim Chown <>
  • Cc: "" <>, Frédéric LOUI <>, Carmen Misa Moreira <>
  • Subject: Re: [rare-dev] Prometheus DB and Grafana Dashboards
  • Date: Tue, 7 Dec 2021 14:08:14 +0100



On 12/7/21 13:53, Tim Chown wrote:

That s to be defined. We reported a little on streaming telemetry in a
Task 3 deliverable / white paper earlier in the project, and when we had the
WP7 network evolution studies it was a topic mentioned by many NRENs, so
there s clear interest. What we need to do is get details of the specific
interest and what we could do to investigate that further.

An example is our current tests of data transfers to Brazil. We can push
80Gbit/s+ on the 100G BELLA link, but because GEANT s new BRIAN monitoring
system is classic SNMP 5-min polling, you don t see that peak unless you
test for 10 mins to guarantee you fill a 5-min slot.

So the question is perhaps where does the instantaneous measurements/data
help you.


so interface counters is a thing.... but streaming telemetry (with occasional
push) can do much more:
for example one can stream (at least from rare/freertr)
- bgp states (up/down) (even from an implicit source such as bmp!)
- igp topology/link/whatever changes (in this case freertr must be a router
node)
to influx/grafana/whatever, where from, you can alert within seconds!
imho it should ring the bell?! :)

thanks,
cs



Archive powered by MHonArc 2.6.19.

Top of Page