EC2 instances usage should only count 'running' instances

[pdecat]

Hi,

I believe the Running On-Demand EC2 instances AWS service limits only count EC2 instances in the running state.

I could not find any explicit statements on that subject in the AWS documentation but I have got the case where my instances count is above the service limit:

# awslimitchecker --list-limits | grep "Running On-Demand EC2 instances"
EC2/Running On-Demand EC2 instances                    200 (API)

# awslimitchecker --show-usage | grep "Running On-Demand EC2 instances"
EC2/Running On-Demand EC2 instances                    207

Filtering on the instance state, I currently have:

• 170 running instances
• 22 stopped instances
• 11 terminated instances

[jantman]

@pdecat You're right, sorry. I've missed that this whole time (I guess because none of my employers leave many stopped instances around). The relevant code is https://github.com/jantman/awslimitchecker/blob/master/awslimitchecker/services/ec2.py#L83

I'll try to get a fix for this tonight or tomorrow. Thanks for taking the time to make me aware of this!

[jantman]

@pdecat I'm not sure how I missed this, but the code that I referenced above handles reserved instances; it explicitly queries all active reservations (by AZ and instance type) and subtracts that from the on-demand instances. It does not, however, ignore instances in the "stopped" or "terminated" instance state.

I've got a fix for this and will be making a bugfix release shortly.