Quick question, I’ve been using the inputs.cisco_telemetry_mdt input plugin to capture streaming telemetry being sent to the Telegraf server from Cisco switches. So far it seems to be working well, but saw the gnmi plugin and noticed that the community is starting to gravitate that way, so I went ahead and checked it out.
I had question:
The gnmi plugin appears to be a pull model, where the Telegraf agent has a set of credentials, subscriptions, and switch list that it uses to poll the switches in order to get the telemetry to pass to serializers/output plugins. Does this support a push method like the inputs.cisco_telemetry_mdt plugin? That way I can configure the switches to send the gRPC data to Telegraf and not the other way around, and also allows for redundancy behind a load balancer.
I recommend checking out this recording from the Cisco team members that built this plugin. If you have questions after watching this, I can connect you to them.
Hi,
This video was packed with great information but it wasn’t clear to me how to achieve redundancy.
The presenter briefly touched on this but didn’t explain how to achieve redundancy when using dial in, which appears to be how the plugin works.
Currently with the cisco_telemetry_mdt plugin I am using the dial out, where the Cisco switch sends data to the Telegraf container that’s sitting behind a load balancer.
Does the plugin support both ways, and if not, what is best practice for achieving redundancy whilst using dial in?
I don’t know about the community at large, but we aren’t planning on removing support for the cisco_telemetry_mdt plugin, so you’re free to continue using it.
It might be worth opening an issue to ask for docs clarification on this question and I could try to get the original author to comment.
