As a guide for others there are two traps I fell into:
- I did not realise AWS has default limits on the number of instances of certain types. In creating multiple CALs I exceeded one of these limits. AWS has a form where you can increase this limit.
- I did not realise that terminating an instance does not clean out the old elastic IP addresses and so quickly exceeded the default limit of five. Later I cleaned up my unused IPs and also increased my limit.