**Service limits and quotas* - Your call rate to an AWS service might be too frequent, or you might have reached a specific AWS service quota. In either case, without proper error handling you wouldn’t know or wouldn’t handle them.
**Parameter validation and checking* - API requirements can change, especially across API versions. Catching these errors helps to identify if there’s an issue with the parameters you provide to any given API call.
**Proper logging and messaging* - Catching errors and exceptions means you can log them. This can be instrumental in troubleshooting any code you write when interacting with AWS services.
These exceptions are statically defined within the botocore package, a dependency of Boto3. The exceptions are related to issues with client-side behaviors, configurations, or validations. You can generate a list of the statically defined botocore exceptions using the following code:
AWS service exceptions are caught with the underlying botocore exception, ``ClientError``. After you catch this exception, you can parse through the response for specifics around that error, including the service-specific exception. Exceptions and errors from AWS services vary widely. You can quickly get a list of an AWS service’s exceptions using Boto3.
For a complete list of error responses from the services you’re using, consult the individual service’s `AWS documentation <https://docs.aws.amazon.com/>`_, specifically the error response section of the AWS service’s API reference. These references also provide context around the exceptions and errors.
Botocore exceptions are statically defined in the botocore package. Any Boto3 clients you create will use these same statically defined exception classes. The most common botocore exception you’ll encounter is ``ClientError``. This is a general exception when an error response is provided by an AWS service to your Boto3 client’s request.
Additional client-side issues with SSL negotiation, client misconfiguration, or AWS service validation errors will also throw botocore exceptions. Here’s a generic example of how you might catch botocore exceptions.
..code-block:: python
import botocore
import boto3
client = boto3.client('aws_service_name')
try:
client.some_api_call(SomeParam='some_param')
except botocore.exceptions.ClientError as error:
# Put your error handling logic here
raise error
except botocore.exceptions.ParamValidationError as error:
raise ValueError('The parameters you provided are incorrect: {}'.format(error))
Parsing error responses and catching exceptions from AWS services
Unlike botocore exceptions, AWS service exceptions aren't statically defined in Boto3. This is due to errors and exceptions from AWS services varying widely and being subject to change. To properly catch an exception from an AWS service, you must parse the error response from the service. The error response provided to your client from the AWS service follows a common structure and is minimally processed and not obfuscated by Boto3.
Using Boto3, the error response from an AWS service will look similar to a success response, except that an ``Error`` nested dictionary will appear with the ``ResponseMetadata`` nested dictionary. Here is an example of what an error response might look like::
Boto3 classifies all AWS service errors and exceptions as ``ClientError`` exceptions. When attempting to catch AWS service exceptions, one way is to catch ``ClientError`` and then parse the error response for the AWS service-specific exception.
Using Amazon Kinesis as an example service, you can use Boto3 to catch the exception ``LimitExceededException`` and insert your own logging message when your code experiences request throttling from the AWS service.
Additionally, you can also access some of the dynamic service-side exceptions from the client’s exception property. Using the previous example, you would need to modify only the ``except`` clause.
..code-block:: python
except client.exceptions.LimitExceedException as error:
logger.warn('API call limit exceeded; backing off and retrying...')
..note::
Catching exceptions through ``ClientError`` and parsing for error codes is still the best way to catch **all** service-side exceptions and errors.
Catching exceptions when using a resource client
------------------------------------------------
When using ``Resource`` classes to interact with certain AWS services, catching exceptions and errors is a similar experience to using a low-level client.
Parsing for error responses uses the same exact methodology outlined in the low-level client section. Catching exceptions through the client’s ``exceptions`` property is slightly different, as you’ll need to access the client’s ``meta`` property to get to the exceptions.
Using Amazon S3 as an example resource service, you can use the client’s exception property to catch the ``BucketAlreadyExists`` exception. And you can still parse the error response to get the bucket name that's passed in the original request.
..code-block:: python
import botocore
import boto3
client = boto3.resource('s3')
try:
client.create_bucket(BucketName='myTestBucket')
except client.meta.client.exceptions.BucketAlreadyExists as err:
As stated previously in this guide, for details and context around specific AWS service exceptions, see the individual service’s `AWS documentation <https://docs.aws.amazon.com/>`_, specifically the error response section of the AWS service’s API reference.
Botocore exceptions will have detailed error messaging when those exceptions are thrown. These error messages provide details and context around the specific exception thrown. Descriptions of these exceptions can be viewed `here <https://github.com/boto/botocore/blob/develop/botocore/exceptions.py>`_.
Outside of specific error or exception details and messaging, you might want to extract additional metadata from error responses:
**Exception class and error message* - You can use this data to build logic around, or in response to, these errors and exceptions.
**Request ID and HTTP status code* - AWS service exceptions might still be vague or lacking in details. If this occurs, contacting customer support and providing the AWS service name, error, error message, and request ID could allow a support engineer to further look into your issue.
Using a low-level Amazon SQS client, here’s an example of catching a generic or vague exception from the AWS service, and parsing out useful metadata from the error response.