We use the zendesk Incremental ticket event API to extract details of ticket updates.  We supply a start_time timestamp which indicates when we want to get ticket events from, we also side load comments.  A typical request URL might be

https://subdomain.zendesk.com/api/v2/incremental/ticket_events.json?incode=comment_events&start_time=<UTC>

Normally we will set the start time to be within the last few minutes (we are not typically extracting data from long ago).

The problem we have is that occasionally, the resulting data contains events that occurred way before the specified start time.  For example we asked for data from December 13 2022 and we got back events which were timestamped as 3 May 2021.  Moreover, in this case we got a huge number of these 'ancient' events.My understanding is that the incremental ticket event API will return a maximum of 1000 events, but we were returned 7435 of them in a single call and all bar 14 of them were from well before the time period we asked for.

There are a number of issues this causes

• We are seeing and processing invalid data
• Due to the size of the data returned we are hitting timeouts and are unable to process the data at all.

This does not happen all the time, but we have seen it on a number of subdomains and it makes the use of the incremental ticket event API unreliable.  Can something be done to ensure that the events returned are only those which are valid for the time period asked for.

I can provide specific examples of REST API calls that produce this rogue data.