As the title suggests, Nomad is rate-limiting the Nomad Autoscaler when it begins monitoring thousands of services. I understand that increasing [`http_max_conns_per_client`](https://developer.hashicorp.com/nomad/docs/configuration#http_max_conns_per_client) or adjusting [`evaluation_interval`](https://developer.hashicorp.com/nomad/tools/autoscaling/policy#evaluation_interval)/[`default_evaluation_interval`](https://www.nomadproject.io/docs/autoscaling/agent#default_evaluation_interval) can help, but at scale, we eventually run into the same issue. I’m more than happy to contribute a fix for this! 🙂