I created a Mandrill website that will update my application whenever the email address bounces or bounces badly, so I do not store this email address in my database. The way this works is this: the user gives me the address, I send him a confirmation, and if I don’t hear from the Mandrill website in 30 minutes, I assume that everything is in order.
So, I did some tests with nonexistent addresses, and they did not go too well. Most of them appeared within a few hours, after I assumed that everything was in order.
In addition, I did not take into account the delay in receiving web chat packages. One mail bounced at 2:01 pm, according to outgoing activity logs, but a web search history shows that the packet only leaves at 2:52 pm.
My questions are: how long should I put my application on a delivery assumption to give Mandrill enough time to detect hard bounces / deviations, and then send me webhook packages? I can live by letting me skip about 5% of bad emails due to a delay in processing during peak or other unusual events, but it seems to me that my 30 minutes are not enough to catch anything at all ...
source
share