Hangfire is our Messaging Bus platform for a fairly large application (48,000 active users) and sometimes, the Windows services stop randomly, or a downstream service has trouble keeping up with the amount of requests we take a second (anywhere from 40-100).
In any case, when one of these “service delays” occurs, jobs start piling up in Hangfire, and sometimes the queue has reached upwards of 100,000 jobs. We then must fix the issue and watch the jobs drain (which may take several hours depending on the rate of new requests).
Now for my question. Is there any way we can set up a Hangfire process to read in messages last in, first out? This will eliminate the issue of service delays entirely, and process new requests while we figure out what has/is happening downstream.
Thank you for your help!