We use AWS SQS and Rabbit. At our scale, SQS is easy peasy and we can wrapper it to make http calls instead of using SQS, as we're using AWS Beanstalk workers.
SQS is generally quicker to get up and running with and we can have metrics out the box. With rabbit we use it for some other stuff and it works just fine, it's when things go into a black hole we struggle, but that's our lack of knowledge.
Depending on your scale, we find SQS is cheaper than a managed rabbit service.
Although I'd be interested in using kafka!
Depending on your scale, we find SQS is cheaper than a managed rabbit service. Although I'd be interested in using kafka!