At work, my development team is contracted with a company that uses AWS, and for better or worse, we have also become the devops team. We have been burned by AWS before, and we have a rule of thumb: if you are deploying new functionality/service communication, after deploy, monitor for 10-15 minutes, with a wide enough window to see if there is a noticable/unexpected change from before the deploy. It always feels lik…
... have you considered automating this? Alarms are pretty straight forward across all cloud platforms. Since you're using AWS: CloudWatch has anomaly detection. I haven't used it personally but perhaps it's worthwhile to look into: https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitori...