Backend API performance metrics

[shailendrasingh]

Dear Team,

I cam across a requirement of back end API performance metrics. They are asking to provide them latency data. I am unable to understand which latency data I should provide them ?

Could you throw some light on this ?

Regards, Shailendra Singh

[sidd-harth]

I guess you will get your latency data from the Target performance dashboard.

https://docs.apigee.com/api-platform/analytics/endpoint-performance-dashboard

Metric	Description
Response Time	Total number of milliseconds it took to respond to a call. This time includes the Apigee API proxy overhead and your target server time.
Target Response Time	Number of milliseconds it took your target server to respond to a call. This number tells you how your own servers are behaving.
Request Processing Latency	Number of milliseconds from the time when a call reaches the selected API proxy to the time when Apigee sends the call to your target server. Add the request and response latencies to calculate the final overhead the API proxy added to the call.
Response Processing Latency	Number of milliseconds from the time when the API proxy receives your target server’s response to the time when Apigee sends the response to the original caller. Add the request and response latencies to calculate the final overhead the API proxy added to the call.

There is a latency analysis dashboard in few Apigee Plans.

https://docs.apigee.com/api-platform/analytics/latency-analysis-dashboard

===

You can also calculate the Latency in the proxy using JS, and Apigee variables and then push to Custom Analytics.

[shailendrasingh]

Thank you Sid !

However I am not able to understand below latency:

• The internet latency--I am unable to understand this, what is role of this latency ?

• The network latency---I understand this one if you need to calculate performance of backend API you need to factor this latency.

• The backend latency-- I am unable to understand this, what is role of this latency ?

• The Apigee proxy latency---This latency gives you overall API performance.

Regards,

Shailendra Singh