Closed iainbryson closed 4 years ago
Each replica | instance has to check it's own status. example:
If you have a single queue you have to requeue the job until it gets processed by the referencing instance. Previous versions used a single queue approach and were causing a lot of load to redis and very random behavior. https://github.com/arturictus/sidekiq_alive/issues/10
I hope that clarifies why each instance has its own queue.
I'm curious about why each host gets its own queue. As I understand it, Sidekiq recommends as few queues as possible. If I want to put the alives in my
critical
, what problem does that cause?I suppose I can monkey-patch the name, but it seems like this is an 'obvious' feature, so maybe I'm missing some bad implications of doing so.