Throughput analytics at developer portal

[Tomasz_Kowalczy]

Hi Apigee Community,

In my company we have a developer portal connected with on-premise edge installation.

When it comes to the developer portal, we often receive questions about analytics view of customers' usage. The most frequent ones are about Throughput and Endpoint Response Time displays. Can you please give me exact definitions of each display seen in analytics tab?

Thank you in advance,

Tomasz

[#2]

All analytics data comes from Edge Management Server. Take a look at this section of the official documentation, it describes the meaning of all analytics reports and variables:

http://docs.apigee.com/analytics-services/content/analytics-dashboards

[anilsr · Accepted solution]

@Tomasz.Kowalczyk , Great Question, Thanks for asking same,

Analytics labels in Developer Portal is little confusing since they map to metrics with a different name. Please find the mapping below & more details related to same.

return array(
  'total_response_time' => t('Throughput'),
  'max_response_time' => t('Max Response Time'),
  'min_response_time' => t('Min Response Time'),
  'end_point_response_time' => t('Endpoint Response Time'),
  'message_count' => t('Message Count'),
  'error_count' => t('Error Count'),
);

Throughput => Total Response Time,

The amount of time (sum, average, minimum, or maximum), in milliseconds, between when a client makes a request and receives a response. The time includes network overhead (such as the time it takes load balancers and routers to do their work), request processing latency, response processing latency, and target response time (if the response is served from the target service instead of cache).

Using different dimensions, you can examine processing latencies by API proxy, developer app, region, and so on.

API syntax: avg(total_response_time)

End Point Response Time,

I think it's no longer working, It's a bug in Developer Portal Above metric might have been removed from analytics & replaced by target response time but in portal it's not updated at all.

@spadmanabhan , What is Endpoint Response Time - end_point_response_time metric in Edge Analytics ? It's being used in Developer Portal but API call to edge errors out with below message, Can you share any info related to "end_point_response_time" metric ?

{
  "code": "analytics.dataapi.service.ClientError",
  "message": "Invalid API Call - Metric 'end_point_response_time' not found in schema.",
  "contexts": []
}

[Tomasz_Kowalczy]

Thank you, @Anil Sagar for detailed answer.

Please let me know if you have more information about Endpoint Response Time.

[anilsr]

@Tomasz.Kowalczyk ,Glad it helps. We will keep you posted regarding Endpoint Response Time.

Looks like it's older version of target_response_time metric which is no longer working. We will keep you posted.

[anilsr]

@Tomasz.Kowalczyk , Just FYI, You can also use this custom module for better UX & remove unnecessary metric selection dropdowns. I am not sure target_response_time matters for end developer since it's all about API for them & Apigee / Target Server doesn't matter for them.

[#7]

Anil, you are correct about endpoint_response_time. It has been replaced by target_response_time in analytics schema. Pls. send links to any developer portal or other documentation so we can fix this.

[anilsr]

Thank you @padmanabhans , I have created a JIRA ticket on DEVSOL-2569. Developer Portal team has to release this fix. It should be quick 2 min change. @Chris Novak , @harsh FYI ^

@Tomasz.Kowalczyk , FYI, We will keep you posted when fix is released in developer portal. You can also subscribe to release notes.

[Tomasz_Kowalczy]

Thank you, @Anil Sagar.

Regarding the Developer Portal fix, are you referring to the fix to Apigee php sdk?

[anilsr]

@Tomasz.Kowalczyk , Yes, fix to module & sdk. Just thinking big, Why do developer need target response time in developer portal when developer doesn't know apigee sits in between ?

Why does it matter for end developer target vs proxy vs total ? I think just total response time is needed & we should get rid of the target_response_time metric in app analytics.

[#11]

Agree, for a developer, target_response_time won't be useful. Total response is good enough.