Improve Jira Cloud API response time

Question

Hi,

I am using JRJC to fetch issues of a project via the Rest API to perform excel reports. Great java library by the way.

I encounter a very slow response time - less than 1 issue per second. How can I improve this?

My jira client basically

* Execute JQL to retrieve the list of open epics using SearchRestClient
* For each epic, get details using IssueRestClient
* For each epic, get children issues using SearchRestClient
* For each child issue, get details using IssueRestClient

To keep things simple I let the client wait for each query before executing another query :

SearchResult res = searchClient

.searchJql("project=" + project + " and issuetype=Epic")

.get();

Then I get details with issue client :

Issue issue = issueClient.getIssue(epic.getKey()).get();

I need 15 minutes to fetch around 300 issues.

I am surprised because Jira Cloud itself has no problem displaying the project content.

Would my request be faster if I dynamically created a filter? I presume filter are similar to SQL views or index behind the scene.

JRJC version : 2.0.0-m2

Ligia Merciu · Answer

@Martin Pernollet did you get some useful information from Atlassian on this topic? Please, share with us.

@Alexey Matveev Did you find the time to write down that article? Please share!

Thanks, guys!

Martin Pernollet · Answer

Thanks you - this sounds to be a good way to improve response time.

For the moment I am discussing with Atlassian support about this 1 second delay - I will let you know their proposals to have a better performance.

I am definitely interested by your article!

Alexey Matveev · Answer

I also have about 1 sec delay in a single thread. And I wish it were faster. For example, I found out that getting worklogs by the Tempo Rest Api works faster than getting worklogs by the Atalssian Rest Api. Maybe Tempo servers are less loaded than Atlassian servers or there can be other reasons.

But basically I use multiple threads. I do development with the Java language. I wrote abstract classes to handle multithreading by fork join, and then I easily connect any rest call to this abstract classes to get multithreading functionality.

I want to write an article here about it if I have time.