I have a workflow that would be -very- easy to parallelize, except that it makes calls to two REST APIs (Google Gmail and Streak) that have usage throttles. If I “fan out” a dozen instances of the workflow at a time, I’m bound to hit the throttle limits and all of the instances just contend with each other.
What I need is some middleware that can proxy all of these requests to Streak and Google and service all of the API calls from all of the instances as fast as possible within their usage limits, and can handle any pacing and back-off needed.
Any suggestions? I’m a big user of AWS so something in the AWS family would be simple. I’m also a fan of hookdeck and it looks like it’ll work for this, but hookdeck does its throttling separately for each endpoint where the services have quotas have usage quotas that span all endpoints, so ideally I’d find something that can handle usage quotas that span all endpoints from a service.
Thanks,
Lee