- 최신
- 최다 투표
- 가장 많은 댓글
Do you use any other services? Like S3? And your EC2 instances are communicating with S3 buckets in us-east-1 region? If so, make sure that you're using Gateway VPC endpoints for S3.
Otherwise, when you go to Cost Explorer, and apply filter "Usage Type" - and search for %Regional-Bytes%, and then Group your result by Service. Are these charges associated only/mostly with EC2?
Billing tools (even most detailed AWS Cost and Usage Reports with resource IDs) only can show information about resources that generate the charges, however they won't be able to show exact information about source or destination of the traffic depending on the traffic type. For example, for data transfer OUT you can see the resource IDs (in CUR) that traffic is coming from, but you can't see where exactly that traffic goes; and for data transfer IN you can see the resource that's accepting ingress traffic, but you can't see which exact resource it's coming from.
[ADDITION] - if all instances are in the same AZ, but they are using Public IP (like Elastic IP) to communicate, this will be the most likely cause of generated DataTransfer-Regional-Bytes charges. Instances in the same AZ that communicate within private network (using private IPs) within AWS, have no data trasnfer charges. However, if with the same setup you use public IP, you will see data transfer charges associated with those resources. Check in Cost Explorer - filter DataTransfer-Regional-Bytes usage type, and group by API Operation. The result might show operations like "public IP" or something similar in this case.
If this is mostly for EC2, your best bet could be using VPC Flow Logs - they may actually show the source/destination of traffic, including ports, IPs, etc. Note though, that flow logs can get quite expensive if you leave them turned on for long period or time and you are tranferring large volumes of data, and they also won't show what happened in the past before you turned them on. If you don't use them regulrly, you can turn them on for a period of time for troubleshooting, and turn them off later.
I have only one instance on EC2 that communicates with S3. But the volume of data is above the previous ones, is there a way to limit this daily transfer so as not to exceed 10 million?
Thanks for the reply. The only services am using are EC2, no S3. Cost explorer grouped by Service + Regional-bytes filter shows 100% of the data/charges as EC2-other. I'll give VPC Flow Logs a shot.
Ah, one more thing - if all instances are in the same AZ, are you using Public IP on them to communicate with each other? Like... Elastic IP for example? If so - this will be the cause of generated DataTransfer-Regional-Bytes charges. If instances are in the same AZ, and they communicate within private network (using private IPs) within AWS, then that traffic is free. However, if with the same setup you use public IP, you will likely see data transfer charges associated with those resources. Check one more thing in Cost Explorer - when you filter DataTransfer-Regional-Bytes usage type, can you group by API Operation? If the results will show the operations like "public IP" or something similar, then my last theory is right :)
That's it, it's all PublicIP-In / PublicIP-Out usage. It wasn't obvious to me from the documentation so thanks for pointing that out.
Sorry to ask in this post, if our ECS/EC2 uses RDS Endpoint, does it count as a private or public connection?