API Proxy Performance

This page applies to Apigee and Apigee hybrid.

View Apigee Edge documentation.

What does this dashboard tell me?

The Proxy Performance dashboard helps you see API proxy traffic patterns and processing times. You can easily visualize how much traffic your APIs generate and how long it takes for API calls to be processed, from the time they are received by Apigee until they are returned to the client app.

The Proxy Performance dashboard

Access the Proxy Performance dashboard, as described below.

To access the Proxy Performance dashboard:

If you are using the Apigee UI in Cloud console: Select Analytics > API metrics and click the API Proxy Performance tab.
If you are using the classic Apigee UI: Select Analyze > API Metrics > API Proxy Performance.

The API Metrics dashboard opens with the API Proxy Performance tab selected, as shown below:

What does this dashboard measure?

This dashboard includes these charts:

Traffic
Average Response Time
Traffic by Proxy
Average Response Time by Proxy

Note: This client.received.start.timestamp time value is used to calculate traffic metrics for API proxies in the analytics dashboards. For more information, see the client flow variable reference.

Traffic

Metric	Description
Total Traffic	The total number of API requests received by Apigee for an API environment in an organization.
Traffic Success	The total number of requests that resulted in a successful response. Error responses do not count.
Traffic Errors	The total number of all API requests that are unsuccessful, that is, the request does not deliver a response as desired by the end user. The count includes both Proxy and Target errors.
Average TPS	The average number of API requests and resulting responses per second.

Average Response Time

See also this interesting article on the Apigee Community site: When can the Average Total Response Time be less than the Average Target Response Time?

Metric	Description
Average Response Time	The average of the Total Response Time measured for all API calls made to an Apigee organization environment. The Total Response Time is the amount of time it takes for an API call to Apigee to return (in milliseconds). Or, put another way, total response time is the time measured from when an entire API call is received on Apigee to the time Apigee begins sending a response back to the client app. This chart measures the average for all proxies. For individual proxies, see the Average Response Time by Proxy chart below.
Average Proxy Response Time	This value is calculated as the average of the Total Response Time minus the Target Response Time for all API calls made to an Apigee organization environment. It's basically a measure of how much time the API calls spend flowing through Apigee itself (in milliseconds).
Average Target Response Time	The average number of milliseconds that it takes from the point the last byte of a request is sent from Apigee to a backend target to the time Apigee receives the last byte of the response. It's basically measuring how much time the API call spends on the target system.

Metric

Description

Average Response Time

The average of the Total Response Time measured for all API calls made to an Apigee organization environment. The Total Response Time is the amount of time it takes for an API call to Apigee to return (in milliseconds).

Or, put another way, total response time is the time measured from when an entire API call is received on Apigee to the time Apigee begins sending a response back to the client app.

This chart measures the average for all proxies. For individual proxies, see the Average Response Time by Proxy chart below.

Average Proxy Response Time

This value is calculated as the average of the Total Response Time minus the Target Response Time for all API calls made to an Apigee organization environment.

It's basically a measure of how much time the API calls spend flowing through Apigee itself (in milliseconds).

Average Target Response Time

The average number of milliseconds that it takes from the point the last byte of a request is sent from Apigee to a backend target to the time Apigee receives the last byte of the response.

It's basically measuring how much time the API call spends on the target system.

Average proxy response time is calculated as the difference between the averages of total response time and target response time. The expectation is that total response time is always greater than the target response time. However, there are also calls which are handled by the proxy itself - preflight requests, cached responses, Oauth tokens etc which do not reach the backend system. These "faster" calls by the proxy reduce the average Total Response time without reducing the Target response time. Hence the average total response time may not always be greater than the average target response time which could result in negative values being seen for proxy response time. When you look at larger TPS without zooming in the time scale too much, these exceptions will not hold much weight and hence give you more accurate and actionable data.

You obtain the most meaningful average response data when a large number of requests are collected over a long period of time, rather than from just a few requests collected over a short period of time. If you only observe a small number of requests that come in over a short time period, there's a chance that even one call that does not go to the target will result in a negative average proxy response time. When evaluating average response times, it's best to view longer time frames that have high traffic.

Traffic by Proxy

Metric	Description
*<Proxy name>*	For the specified API proxy, the number of recorded API requests and responses.

Average Response Time by Proxy

Metric	Description
*<Proxy name>*	For the specified API proxy, the average of the Total Response Time measured for all API calls made to an Apigee environment. The Total Response Time is the amount of time it takes for an API call to Apigee to return (in milliseconds). Hover over the graph to see the total amount of time spent on the proxy side and the target side, as well as the average.