[j-nsp] Junos Telemetry Interface (JTI)

Wed Oct 16 12:05:39 EDT 2019

Response to old thread and referencing my new thread… 

“[j-nsp] telemetry analytics - mx960 - npu packet rate concerns”

My JTI/OpenNTI systems seems to be continuing to run nicely.  I use it when I really need to know some details.

In the new thread I mentioned above, I was curious about my CGNat MS-MPC-128G resource utilization and starting digging around in those sensor agent and their data I’ve been receiving in my telem station.

I’ll copy and paste from that other thread for convienience.

Using my JTI/Chronograf/Grafana web interface I'm trying to understand some of the telemetry analytics data I'm seeing coming from what appears to be the sensor resource of my MX960 corresponding to /junos/system/linecard/npu/utilization/ .. The field seen on chronograf that I'm watching is "npu_util_stats.packets.rate"

When using the Chronograf data explorer and picking one MX960 and a certain _seq number (0-14 , I don't know what these are) I'm seeing some significant drops in the graph during peak time (approx. 7 - 10 p.m.) watching "npu_util_stats.packets.rate" with mean function (as opposed to median, count, min, max, etc, etc).  In other words, the graph shows a typical ramp-up approaching peak times, and ramp-down during the late night hours..

But about a week ago, I started seeing dramatic drops/sags in the graph during those 7-10 p.m. hours.

I'd like to try to figure out what those drops are related to. I'm wondering if this is the MS-MPC-128G npu's in-use for my cgnat.. I've been loading it up quite a bit lately with thousands more subscribers behind it, and am trying to watch how it scales. and if I have any reason for concerns regarding resource load, etc.

If you unicast email me, I’ll send you screen shots of my telem graphs that are giving me concern and the sensor agent strings that I’m looking at.

Thanks y’all

- Aaron

From: Colton Conor [mailto:colton.conor at gmail.com] 
Sent: Monday, November 12, 2018 6:47 AM
To: Aaron
Cc: beecher at beecher.cc; Juniper List; james.burnett at geant.org
Subject: Re: [j-nsp] Junos Telemetry Interface (JTI)

Guys,

I wanted to follow up and see how things are going with JTI?

Also, it has been brought to my attention that OpenNMS supports JTI. I was not aware of that, so I figured I would share with others: https://docs.opennms.org/opennms/branches/develop/guide-admin/guide-admin.html#ga-telemetryd

On Thu, Oct 11, 2018 at 12:24 PM Aaron1 <aaron1 at gvtc.com> wrote:

Yes Niall, lets stay in touch.

Thanks Tom, I’ll have to look at Panoptes

Aaron

> On Oct 11, 2018, at 8:18 AM, Tom Beecher <beecher at beecher.cc> wrote:
> 
> Related, my company open sourced a tool we've been working on for network telemetry at NANOG in Vancouver. I'm 95% sure that a JTI receiver is functional on our internal builds, but they're still working on a few things with streaming receivers generally, so it's not yet in the public repo. May be something that can meet your needs at some point if you wanted to keep an eye on it. 
> 
> https://github.com/yahoo/panoptes
> 
>> On Thu, Oct 11, 2018 at 9:02 AM Niall Donaghy <niall.donaghy at geant.org> wrote:
>> Fantastic news Aaron!
>> 
>> That tallies with our experience of deploying the 'bundle' version of OpenNTI 
>> for Junos ST.
>> 
>> We look forward to your shared experiences as you kick the tyres and - 
>> hopefully - incorporate this into your NMS/procedures. :)
>> 
>> Many thanks,
>> Niall
>> 
>> 
>> -----Original Message-----
>> From: Aaron Gould [mailto:aaron1 at gvtc.com]
>> Sent: 11 October 2018 13:59
>> To: juniper-nsp at puck.nether.net
>> Cc: James Burnett <james.burnett at geant.org>; Niall Donaghy 
>> <niall.donaghy at geant.org>; 'Colton Conor' <colton.conor at gmail.com>
>> Subject: RE: [j-nsp] Junos Telemetry Interface (JTI)
>> 
>> Wanted to circle back with y'all... I finally got this working...thanks to 
>> techmocha10 (see below) and my linux coworker genius (Dave),
>> 
>> I'll just copy/paste a post I just made...
>> 
>> https://forums.juniper.net/t5/vMX/Telemetry-data-is-not-streaming-from-Juniper-vMX-17-4R1-16/m-p/375996#M923
>> 
>> 
>> I got telemetry streaming working using this site ... I have a couple MX960's 
>> streaming telemetry to the suite of software provided in this Open-NTI project 
>> spoken of on this techmocha blog site.  I think my previous problems were 
>> related to conflicting installs.... as myself and my coworker had loaded 
>> individual items and then the open-nti suite (which i understand is a docker 
>> container with all the items like grafana, fluentd, chronograf, influxdb, 
>> etc).... anyway, we started with a fresh install Ubunto virtual machine and 
>> *only* loaded Open-NTI and it works.
>> 
>> 
>> I do not know or understand all of the innerworkings of it at this point, but 
>> am quickly learning, even while writing this post... I'm currently using 
>> Chronograf hosted at port 8888 and browsing the Data Explorer function and 
>> seeing some nice graphs.  (I'm wondering if Chrongraf is simply an alternative 
>> to Grafana gui front end, unsure) There seems to be tons of items to monitor 
>> and analyze, and I'm currently only sending the following sensor resource from 
>> the MX960 and there are several more that can be sent.... 
>> /junos/system/linecard/interface/
>> 
>> 
>> I am sending the telemetry from the MX960 using UDP transport and GPB format 
>> to port 50000 and source port 21111 (mx960-1) and 21112 (mx960-2).  I'm unsure 
>> that I had to use unique source ports... as I wonder if the source-ip would 
>> have been sufficient to make the streaming sources unique in the Open-NTI 
>> server.
>> 
>> 
>> Looking at the techmocha pictures, and the "docker ps" command on the linux 
>> server, and now this new-found techmocha link (see "deconstructed" below) 
>> apparently FluentD is the TSDB (time series db) that is receiving/ingesting 
>> the *Native* streaming form of telemetry from my MX960's on udp port 50000 and 
>> looks like fluentd hands off that data to InfluxDB port 8086 (which i think 
>> happens internally at that server).  (I'm not evening talking about the other 
>> form of jti telemetry using openconfig and grpc....I've yet to do that and 
>> don't know why I would exactly...which i beleive is ingested using telegraf, 
>> unsure)
>> 
>> 
>> ...the link i followed to deploy open-nti suite....
>> https://techmocha.blog/2017/06/26/using-opennti-as-a-collector-for-streaming-telemetry-from-juniper-devices-part-1/#comments
>> 
>> 
>> ...interestingly, i just now found this, which apparently is a way of 
>> deploying all the components individually...
>> https://techmocha.blog/2017/10/31/serving-up-opennti-deconstructed/
>> 
>> 
>> 
>> _______________________________________________
>> juniper-nsp mailing list juniper-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp