Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Performance Monitoring

One primary function of the Junos Telemetry Interface is performance monitoring. Streaming data to a performance management system enables network administrators to measure trends in link and node utilization, and troubleshoot such issues as network congestion in real time.

In a typical deployment, the network element, or device, streams duplicate data to two destination servers that function as performance management system collectors. Streaming data to two collectors provides redundancy. See Figure 1 for an illustration of how the performance management system collectors request data and how the device streams data. The device provisions sensors to collect and export data using command-line interface (CLI), configuration through NETCONF, or gRPC subscription calls. The collectors request data by initiating a telemetry subscription. Data is requested only once and is streamed periodically.

Figure 1: Telemetry Streaming for Performance Management Diagram of a network telemetry system with devices connected to a performance management application. Blue dashed lines show telemetry data streaming; orange dashed lines show telemetry subscription updates using NETCONF, CLI, OpenConfig RPC.

Starting in Junos OS Release 18.1R1, a new sensor is available that allows syslog data to be streamed to network telemetry collector systems. Using the /junos/events/ sensor, and an export profile with a reporting-rate of 0, you can now stream event data along with statistical data to your telemetry-collection systems.

Other applications of the Junos Telemetry Interface include providing real-time data to support operational state synchronization between a network element and an external controller, such as the Northstar Controller, which automates the creation of traffic-engineering paths across the network. The NorthStar Controller can subscribe to telemetry data about certain network elements, such as label-switched path (LSP) statistics.