At Move Work Forward, we strive to provide the best product we can, considering the limitations we've got. Each year we need to recertify our Atlassian Apps for the Data Center support of Jira, Confluence and Bitbucket.
In this post, we share with you more details on how we were testing one of our Atlassian Jira App - Jigit - GitHub and GitLab integration for Jira:
We built our production clusters in Amazon Web Service with the help of Atlassian-provided automation scripts. In the same region, we run our clients to load Jira and installed applications.
The main focus of our Jira App is to show GitHub or GitLab comments as part of the Jira issue view, which requires constant scanning of the connected GitHub/Gitlab repositories. When the operation starts, the indexing may take substantial time, depending on the size of the repository or repositories group to work with.
For our performance test, we picked a huge repository - nearly 30 000 commits and nearly 500Mb is size.
Below you can find the instance information.
Here is the AWS instance information of one of our Jira nodes.
Below you can see Jira Runtime information.
One of the tests we have done is to see how GitLab and Jira are loaded under this initial indexing process. Below you can see the monitoring information about GitLab AWS instance.
The Jira database is monitored.
Jigit App, GitHub and Gitlab integration for Jira can hammer down the GitLab or GitHub instance if not configured correctly. Below you can see the test configuration.
As you can see, it is configured to make ten requests every 10 seconds, which should not exceed 3600 requests per hour. It is imperative if you are running Jigit against GitHub or GitLab in the cloud and not Enterprise offering, which may lead to the throttled requests because those services may consider Jigit App as a DDOS attacker.