Intermittent status 403 response for valid REST API calls
A customer instance is seeing their issue creation(and other REST API calls) fail occasionally. Tracing this back it seems to be caused by a call failing and returning a 403. This seems to be an infrequent event as the same calls succeed other times.
Steps to Reproduce
Send a REST API call
We're POSTing to https://<instance>.atlassian.net/rest/api/2/issue?<JWTMASKED> and getting a 403. The response body is not the usual JSON you get from a usual 403 error.
First noticed on 10 Jun 2019. It behaved fine for a while and another spike in 403s was experienced on 13 Jun 2019
Checked if this is related to rate-limiting and was confirmed by the SRE team that the said instance did not have a rate-limiting event
The trace ID for the request is ATL-TraceId: 2b33c88f62003df2 (June 10) and ATL-TraceId: 2d5d538eb8d42a8a (June 13)
I also described my case with 403 here:
I am having a nearly identical problem to Rob Williams.
I have automation rules that apply to parents and sub-tasks.
When a rule applies to, and takes action against, multiple sub-tasks, often there is one or more sub-task missed by the rule. 403 error returned by API.
This error is also preventing me from rolling out an enterprise-wide employee ticketing system.
User here – it has happened in our testing using rules to (first) create a bunch of subtasks in a row (about 38) and (second) to transition a bunch of subtasks. It’s happened 3 times in perhaps 20 cycles of testing on the create/transition logic, which seems not uncommon. Happy to provide more info if it helps – this is holding up a launch to use Jira as a task mgmt system for us, and we may have to scrap the idea without a fix!