When the ray cluster autoscales on AWS the serve replica is always in a state of pending creation on the new ec2 instance.
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Autoscaling Replicas in Ray Serve | 5 | 1660 | March 12, 2021 | |
[Autoscaler] [Clusters] Understanding Ray Autoscaler with Transient Hardware | 2 | 356 | April 5, 2021 | |
Ray serve deployment is not scaling up, ongoing request is always 0 | 1 | 298 | April 18, 2024 | |
Ray Clusters with AWS IAM roles | 1 | 224 | January 14, 2025 | |
Ray Serve Autoscaling: Autoscaling backend-replicas removed? | 3 | 491 | February 18, 2021 |