Right now, we can't terminate queued jobs if there are more than 10 jobs in the queue. It's an issues with the SQS APIs.
So instead, when we run, put the job in the queue and set DB analysis status to queued. When we take it off the queue, switch it to running on the DB and run it locally. When we terminate a queued job in this model, keep reading the job queue until we find that job and take it off.
Think about how the DB and queue could get out of sync...
Right now, we can't terminate queued jobs if there are more than 10 jobs in the queue. It's an issues with the SQS APIs.
So instead, when we run, put the job in the queue and set DB analysis status to queued. When we take it off the queue, switch it to running on the DB and run it locally. When we terminate a queued job in this model, keep reading the job queue until we find that job and take it off.
Think about how the DB and queue could get out of sync...