Fred Birchall

11/17/2022, 9:55 AM
Hi All! I’m wondering if the
block has any future scope to handle Lambda invokes, more specifically long running functions (over 300 seconds). The issue that my team is hitting is due to
TCP keepalive
. We are self hosting our prefect instance using AWS EKS and seen that with some k8s magic (when using EC2 instances) you can redefine the system configuration to send the first TCP keepalive message after say 200 seconds instead of the Linux default of 2 hours, but this solution does not work using AWS Fargate (our current set-up). So I was wondering if this issue is shared with others and if there is a potential solution down the line? Thanks!