Create spot instances in different regions - amazon-ec2

Question

I'm trying to create spot instances in different region using boto3.
My default regions defined as us-east-1.
When I'm trying to create the spot instances in different region from the default one , an exception is thrown with this message:
botocore.exceptions.ClientError: An error occurred (InvalidParameterValue) when calling the RequestSpotInstances operation: Invalid availability zone: [eu-west-2b]
The instances created using this code snippet:
for idx in range(len(regions)):
client.request_spot_instances(
DryRun=False,
SpotPrice=price_bids,
InstanceCount=number_of_instances,
LaunchSpecification=
{
'ImageId': ami_id,
'KeyName': 'matrix',
'SecurityGroupIds': ['sg-5f07f52c'],
'SecurityGroups': ['MatrixSG'],
'InstanceType': machine_type,
'Placement':
{
'AvailabilityZone': regions[idx],
},
},
)

Answer 1

I assume that your regions list actually contains a list of Availability Zones rather than regions (since it says 'AvailabilityZone': regions[idx]).
Each AWS region operates independently. When making a connection to an AWS service, you must connect to the specific service in a specific region.
For example:
client = boto3.client('ec2')
This connects to EC2 in your default region.
Alternatively, you can specify a region:
client = boto3.client('ec2', region_name = 'eu-west-2')
You are receiving the Invalid availability zone error because your client is connected to a region (eg us-east-1) but is referencing an Availability Zone that is in a different region (eg eu-west-2b).
Your code is only creating the client once, but is attempting to connect to multiple regions. The solution is to create a new client connection for the region you wish to use. If your loop connects to multiple regions, then the client should be defined within the loop rather than outside the loop.