Update Kinesis Mock to 0.3.5 #7639
Conversation
See https://github.com/etspaceman/kinesis-mock/releases/tag/0.3.5 In particular, fixes a couple of fields in response types so that they are nullable, rather than empty lists (or 0'd integers).
FailedRecordCount is now null, not 0, for these types of responses
| "stream_description": { | ||
| "StreamDescription": { | ||
| "EncryptionType": "NONE", | ||
| "EnhancedMonitoring": [ |
There was a problem hiding this comment.
These are not returned when empty now. According to the Kinesis API, ShardLevelMetrics must be non-empty if this is returned, otherwise it is null. In this case, it is null.
There was a problem hiding this comment.
Hi @etspaceman !
These are actually validated against AWS. Our *.snapshot.json files are all automatically generated when running against AWS and we generally don't update them manually for this reason. To update them we run the test with the following additional environment variables.
AWS_PROFILE=<your-profile>
TEST_TARGET=AWS_CLOUD
SNAPSHOT_UPDATE=1 # or set it to `0` if you just want to verify that the existing snapshot still validates against AWS
This could've changed since we recorded it, but I just updated to have the newest specs in boto3, re-ran the test and this seems to still be the observed result unfortunately.
Additionally, please note that these are not the raw HTTP responses you're seeing, but the objects generated by the SDK.
There was a problem hiding this comment.
That's very interesting. Here is section in the smithy documents about this:
You can see that the min length here is 1. I've confirmed this in the Java SDK as well.
But also, if I run a describe-stream in the aws-cli, I see empty lists being returned. So I wonder if this is a bug in Smithy.
There was a problem hiding this comment.
@dominikschubert I am going to open a support ticket with AWS on this one to get some clarity. For now I'll close this PR and revert my change in kinesis-mock
| "describe_stream_result": { | ||
| "StreamDescription": { | ||
| "EncryptionType": "NONE", | ||
| "EnhancedMonitoring": [ |
tests/integration/test_apigateway.py
Outdated
| result = requests.post(url, data=json.dumps(test_data)) | ||
| result = json.loads(to_str(result.content)) | ||
| assert 0 == result["FailedRecordCount"] | ||
| assert None == result.get["FailedRecordCount"] |
There was a problem hiding this comment.
This is null now. According to the Kinesis API, FailedRecordCount must be > 0 if it is returned, otherwise it is null.
| # by default, no errors | ||
| test_no_errors = kinesis.put_records(StreamName=stream_name, Records=records) | ||
| assert test_no_errors["FailedRecordCount"] == 0 | ||
| assert test_no_errors.get["FailedRecordCount"] == None |
2 participants
See https://github.com/etspaceman/kinesis-mock/releases/tag/0.3.5
In particular, fixes a couple of fields in response types so that they are nullable, rather than empty lists (or 0'd integers).