using Apigee analytics service for issue detection

[#1]

what is the delay of data collection in Apigee analytics, is there any API to check the aggregation interval?

I want to use the exposed Apigee analytics APIs to do some issue_detection (something like the API HEALTH) in our system.

How much delay is in the Apigee analytics ? does realtime parameter guarantees to get up-to-date analytics data?

So Is it right to use the Apigee Analytics API’s with realtime=true, to check the API-PROXY health stats and get alerted whenever there is a latency anomaly or failures using various available metrics in Apigee analytics like status code, error_count etc ???

[spadmanabhan · Accepted solution]

The data is collected as raw event as soon as possible but the aggregations are run on a schedule (typically 5 mins). So, if your query is processed by an aggregation, it will show a small delay with respect to the raw events. We do not guarantee that the analytics data will be instantly available for querying -- so you should assume and plan for a small delay in your analysis of these events.

You can definitely get alerted on any of these events of interest after querying but assume that this may be a short delay.

The realtime=true parameter will force queries to execute against the event data and hence it willbe more up-to-date but you have to use this carefully by bounding the time range you are looking at.