Hi all
We’re busy building automated UI tests in Confluence to test our Forge Macros we’ve built. We’re running into, what seems to be, queuing issues with the content indexer in Confluence.
As part of our automated tests, and to ensure that anybody in our team can run it, the test pack used to execute the tests, must be created as part of the tests and then deleted after the tests are concluded.
We’ve noticed that when creating new pages in quick succession (or renaming pages for that matter), those pages aren’t always searchable in the page tree search. It may take a minute or two for them to appear. The same behaviour seems to impact the CQL REST Api calls (which I assume also uses the Confluence indexer content)
Using the url of the page, after it was created (or renamed), to navigate to the page works though.
Based on what I’ve tried and seen, staggering the actions seems to help (adding a 2-5 second delay between each create/ edit action). It also helps to change up the sequence of the actions (a few creates, then a few edits, then more creates if needed).
Note that this is all against our sandbox environment and I assume the resources put into scaling is a lot less in sandbox than it is in prod.
Has anybody else encountered these issues? Does anybody have any suggestions on what I can do to make my 1) test pack creation, 2) test execution and 3) cleanup more reliable.
The other option our team has talked about is to rather have a dedicated test pack which is maintained as new development/ tests are built.
For those interested: We’re using Javascript and Playwright for our automated tests.