Why don't you start by filling glaring blind spots in CloudWatch metrics? For example no CloudWatch metrics so no simple monitoring or alerting possible for Fargate Ephemeral storage. Yet if you overfill it, your task can't start since container image gets pulled from ECR and put there at startup. If you set it to max value to compensate for this risk, you pay $175 per year per container if leave running 24/7/365. Start by reviewing your existing CW blind spots please.
very good sharing.could i get a copy of your ppt?
Why don't you start by filling glaring blind spots in CloudWatch metrics? For example no CloudWatch metrics so no simple monitoring or alerting possible for Fargate Ephemeral storage. Yet if you overfill it, your task can't start since container image gets pulled from ECR and put there at startup. If you set it to max value to compensate for this risk, you pay $175 per year per container if leave running 24/7/365. Start by reviewing your existing CW blind spots please.