S3 bucket for logging is created on serverless deploy, and torn down and recreated on re-deploy. This means that old logs are destroyed at re-deploy. On the plus side, this means they get tidied up with the lambda.
If this changes, and probably anyway, we may need some mechanism to clean up old logs to avoid building up vast amounts. (and CloudWatch logs need to be explicitly set to expire and be deleted separately from the lambda they relate to)
S3 bucket for logging is created on serverless deploy, and torn down and recreated on re-deploy. This means that old logs are destroyed at re-deploy. On the plus side, this means they get tidied up with the lambda.
If this changes, and probably anyway, we may need some mechanism to clean up old logs to avoid building up vast amounts. (and CloudWatch logs need to be explicitly set to expire and be deleted separately from the lambda they relate to)