It is extremely important for us to have a way to get email alerts when there is a spike in HTTP error codes (e.g. 5xx, 4xx, etc.). Please provide a way to monitor and send when a load balancer goes above certain thresholds. We had that with DataDog and our custom HAProxy and now that we are moving to your managed load balancers we really miss these alerts. In particular it's important to monitor 5xx errors in an aggregated way.