Getting issues from Jira Cloud - startAt does not work with high values Edited

[Øyvind Rogne]

We have created a script that downloads all issues into our data warehouse. It iterates our issue list using 

https://OURSITE.atlassian.net/rest/api/3/search?jql=project=OURPROJECT_CODE&startAt=<startAt parameter>&maxResults=<startAt + 100>

this works for getting issues, but the script fails when we get to high values for startAt. 

For example 

https://OURSITE.atlassian.net/rest/api/3/search?jql=project=OURPROJECT_CODE&startAt=151600&maxResults=151700

returns the following JSON with a HTTP 500 code

{"errorMessages":["Internal server error"],"errors":{}}

one more thing: I do know for sure that we have more than 160 000 issues. So we should be able to get more than 151700 issues 

EDIT: After doing some investigation, I see that 

startAt=151700&maxResults=151800 works, but startAt=151600&maxResults=151700 does NOT work. This is really strange. In other words, there is a certain offset we are not able to retrieve

[Øyvind Rogne]

one more thing: I do know for sure that we have more than 160 000 issues. So we should be able to get more than 151700 issues