There are certain fields in the database which contain information which is immutable. For example, the worker type that an instance counts for should never change. In the case of any EC2 State Change event, we should try to load the data for that {region, id} row and use it instead of needing to run a describeInstances request.
This is not too high priority but would theoretically cut the number of describeInstances requests we make in half. This would also let us log more information for the cases which result in instance deletion, but should not become mandatory. The reason I want to avoid this requirement is that processing deletions is nearly free, ec2-api-wise.
There are certain fields in the database which contain information which is immutable. For example, the worker type that an instance counts for should never change. In the case of any EC2 State Change event, we should try to load the data for that
{region, id}row and use it instead of needing to run adescribeInstancesrequest.This is not too high priority but would theoretically cut the number of
describeInstancesrequests we make in half. This would also let us log more information for the cases which result in instance deletion, but should not become mandatory. The reason I want to avoid this requirement is that processing deletions is nearly free, ec2-api-wise.