So I have an application where there is 2 events generated async for each request.
1st event indicates the downstream system acknowledged the event and the second event indicates if the initial request succeeded or failed.
90% of the time, the 1st and 2nd event happen within 1-2 seconds of each other 10% the 2nd events can come between 1-3 days and some times not, but we will always have "acknowledgements" right away.
An analogy would be sent emails vs read emails. So 90% of emails are read as soon as they are received and 10% within 3 days and some not read at all.
In our app the other day we noticed that the 90% where read about 5 minutes later when usually it takes 2-3 seconds. So to us it means the system aknowledged the delivery of the message, but for what ever reason they only got delivered 5 minutes later then the usual 2-3 seconds.
Both events have a correlation identifier.
So can ML detects this?