@MarkTheThomas@ircolle@awscloud Nope. No need to scale up instances -> continuous submission to Fargate as soon as jobs are runnable. Batch handles retrying on any throttles automatically.
Metaflow for R was just open-sourced at #useR2020! We are super excited to bring together the best parts of the amazing #rstats ecosystem and modern cloud-based scalability and model operations. Learn more at https://t.co/FDS3lORecD #metaflow#datascience
Major Metaflow release! Deploy your ML workflows in production on a highly available, scalable production scheduler, #AWS Step Functions, without changing anything in your Metaflow code. https://t.co/9fRq1poqns #metaflow@awscloud@NetflixResearch@NetflixOSS
@carkujon@awscloud@AWSSupport Fix is live everywhere. Thanks for your patience. Also, please file us a support case with your CE arn, region, and time you saw impact - we'd like to make sure you're back to normal.
@carkujon@awscloud@AWSSupport We've identified a bug causing this behavior: fix is in flight. As a short-term workaround, disabling and re-enabling the CE will get it back to VALID. Will update when fix is fully deployed.