Application performance monitor
An application performance monitor is a simplified framework that
-
enables you to configure intent-based performance monitors
-
provides real-time, end-to-end application performance visibility filtered by client segments, network segments, and server segments, and
-
collects performance metrics for specific traffic to help optimize application performance.
Key concepts in application performance monitoring
Application performance monitoring includes the following key concepts:
-
Monitoring Profile: A profile is a predefined set of traffic monitors that can be enabled or disabled for a context. As part of this feature, the sdwan-performance profile has been enhanced to include Application Response Time (ART) and media monitors to monitor traffic passing through Cisco Catalyst SD-WAN tunnel interfaces. The sdwan-performance profile has a dedicated policy to filter traffic based on your intent.
-
Context: A context represents a performance monitor policy map that is attached to an interface for ingress and egress traffic. A context contains information about a traffic monitor that has to be enabled. When a context is attached to an interface, two policy-maps are created, one each for ingress and egress traffic. Depending on the direction specified in the traffic monitor, the policy maps are attached in that direction and the traffic is monitored.
-
Traffic Monitoring Specifications: You can choose to filter performance metrics using classification and sampler.
 Note |
A context can be attached to multiple interfaces. Only one context can be attached to an interface. You can modify the context only when it is not attached to an interface. |
Traffic monitoring specifications include the following filters:
-
Classification: Classification is a filter that defines the traffic that should be monitored for specified applications. This filter reduces the load on the device and performance collectors because they only need to monitor performance for specific applications.
-
Sampler: A sampler monitors random traffic flows, based on the sampling rate specified, rather than all the flows. Enabling the sampler reduces scaling and performance impact when the scale of traffic is large.
Application performance monitor provides the following features and benefits:
-
ART can be monitored for TCP flows. Some of the parameters that can be monitored are—server network delay, client network delay, and application delay.
-
Jitter can be monitored for Real-time Transport Protocol (RTP) audio and video traffic.
-
Information about input and output interfaces and local and remote TLOCs can be collected for every flow that matches the performance monitor.
-
Performance monitor can be configured on all WAN tunnel interfaces or specific WAN tunnel interfaces using CLI commands.
-
Global performance sampler is supported. The sampler allows you to monitor random flows based on the sampling rate configured, rather than the entire traffic, therefore, reducing performance and scaling overhead.
The following image shows the performance monitoring workflow:
In this image, performance monitor has been applied globally (on all tunnel interfaces). You also have the option to enable it on specific interfaces. Performance is monitored for traffic going out of, and coming into the WAN tunnel interfaces. Based on the exporter parameters defined in the context that is initiated from the monitoring profile, the metrics that are collected are sent to the third-party collector that is defined. You can then view details of the application or media that you are monitoring using various show commands.
Feedback