@avelit, just wanted to confirm that the limit of 100 changelogs returned by api/2/search and /rest/api/2/issue/ endpoints is now in place. As part of the limitation, we changed the default ordering of changelogs returned by these endpoints from ascending to descending. The reason being that for limited number of items, we want to always provide the most relevant (ie. the most recent) issue changes.
However, the /rest/api/2/issue/{issueIdOrKey}/changelog endpoint that returns all issue changelogs in a paginated list, is ordered as ascending (oldest first).
@istankiewicz, thanks for valuable information. we should change some code due to this information because now 2 rest-ends uncompatible due to different orders. We expected read first 100 with expand=changelog and to add all the rest but now we should use only 1 endpoint.
Could you please clarify, if /rest/api/2/issue/{issueIdOrKey}/changelog endpoint is going to be available in Jira Server too? I can’t see it in latest 5.6 documentation at least: JIRA 7.6.0-m0125
It seems like some sociopath did this since first you get a part of changelog from the issue itself in one order - and if you need more records - you have to get it from another endpoint in different order and thus you need to
get last /changelog page based on “total” field (which I’m afraid sometimes is not guaranteed based on some docs I’ve read)
Upload it to memory and sort the resulting array
Add it to initial changelog records list
While adding ability to sort the result could make it million times easier for the user. I hope someone will eventually think about how ridiculous this looks like.