Almost all popular job processing libraries in all major frameworks support executing 1 off tasks with robust retry policies and even support recurring tasks.
The operational complexity is really just running the sidekiq or celery command in another Docker container or if you don't use Docker, then setting up 1 extra systemd unit file.
I once had a Celery / Flask app running on a $20 / month digitalocean server. It handled over 3 million background jobs per month and it also hosted my DB server, cache server and web front end.
On your pricing page, your highest tier supports up to 100,000 requests per month at $129/month. How much would it cost to do 3 million requests per month?
Also how would you set up custom retry policies?