-
Bug
-
Resolution: Unresolved
-
Medium
-
Severity 3 - Minor
Issue Summary
When using the new experimental endpoint Search for users within an organization that was released for the Organizations REST API, frequent timeouts can occur.
Steps to Reproduce
Send a POST with a JSON payload that combines the emailUsernames and emailDomains parameters to search for a user from a specific domain:
{
"emailUsernames": {
"eq": ["john.doe"]
},
"emailDomains": {
"eq": [
"mydomain.com"
]
},
"expand": ["NAME", "EMAIL"],
"limit": 20
}
Expected Results
The user should be returned without any issues.
Actual Results
While the query usually works, very frequently a 504 gateway timeout occurs after 15 seconds.
The issue happens more frequently when using "eq". When using "contains", the timeouts are not that frequent but they do take some time (around 9s) to be processed, depending on the amount of users in the user base.
Workaround
Please retry the API call or implement a retry mechanism to handle the timeout if you are calling this endpoint from a script.
[AX-210] Endpoint "Search for users within an organization" returns intermittent 504 gateway timeout
Index is missing for the search and I'll add a quick ticket in our backlog for the fix.
The timeouts are as the query takes longer than the gateway timeout. Query for a name should not take 15secs, so you want to check if you can use some DB index (maybe?) to speed up the search. After found first time, the element is in some cache and is delivered within < 1sec. But not if you search for a new name.
Seems there was a change, but the search still takes 10 secs (if searched as in the description), which is too long for automation scripts. The index seems to be still missing.
The API is great if you search for an Atlassian ID, giving a lot of information needed by ORG admin / automation.
But we need a well-performing search for email addresses:
So I really do not know who to reliably search for a users in the ORG if I only have the email address. Please fix this as announced.