Journal Service entries are processed an all nodes, is there any alternative to keep a durable FIFO queue that can be processed on a single node?
Could you post some code or a better example as I have to guess a lot of stuff here? Exactly showing where you are stuck.
What I would like to know is if there is a durable queue that Confluence maintains and is persistent. The usage would be to store actions to later process them in a Confluence job.
AFAIK, Journaling service is the only one service, which provides persistent queue of actions in Confluence. However, you may re-use scheduler service to schedule future jobs to process your data. Scheduler service persists all job configs, so they survive Confluence restarts. Job can be scheduled to run once on all cluster nodes (pls. see JobConfig for more details). The only one problem is that if Confluence node is down at the time of scheduled execution, your job probably won’t be executed after Confluence is back up. Also if you spam scheduler service with too many requests, it may slow down the Confluence.
Hope that helps.