By using AWS re:Post, you agree to the Terms of Use

My AWS Batch job service uses a lot of space to S3, while my code doesn't work with S3?


I have 5 Batch jobs running on AWS Batch with Fargate, when it was running I noticed the capacity to S3 spiked through the NAT Gateway. I queried VPC Logs using Athena and found that the destination IP is of S3 None of my code uses S3, when I turn them off the capacity going to S3 is completely reduced. I don't understand why my Batch job service uses S3 while my code doesn't. Is there any way to investigate to know exactly where the capacity is coming from? (except , I read it). I understand it is possible to use S3 VPC Endpoint to handle throughput through the NAT Gateway, but I want to find the root cause.

1 Answers
Accepted Answer

Your container image (which I assume is in ECR) is stored in s3 which is the likely cause of the s3 traffic spike. An s3 gateway endpoint can be setup to optimise the download - see here

answered 17 days ago
reviewed 16 days ago
  • Ok many thanks, exactly I am using ECR but I remember I saw somewhere on the internet wrote that ECR is using S3 to store images, where can I find evidence of this (eg an AWS blog, tutorial link, ...). I need to prove something...

  • From the link I posted above - "When your containers download images from Amazon ECR, they must access Amazon ECR to get the image manifest and then Amazon S3 to download the actual image layers"

  • Thank you so much

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions