About Amazon SageMaker Ground Truth save in S3 after label - python-3.x

Question

I am setting up the automatic data labelling pipeline for my colleague.
First, I define the ground truth request based on API (bucket, manifests, etc).
Second, I create this labelling job, and all files are uploaded in S3 immediately.
After that my colleague will receive an email saying it is ready to label it, then he will label the data and submit.
Until now, everything is well and quick. Then I check the SageMaker labelling job dashboard, it shows the task is in progress, and it takes very very long time to know it is completed or failed. I don't know the reason. Yesterday, it saved the results at 4 am, took around 6 hours. But if I create label job on website instead of sending requests, it will save the results quickly.
Can anyone explain it? Or maybe I need to set up a time sync or other configuration?
This is my config:
{
"InputConfig": {
"DataSource": {
"S3DataSource": {
"ManifestS3Uri": ""s3://{bucket_name}/{JOB_ID}/{manifest_name}-{JOB_ID}.manifest""
}
},
"DataAttributes": {
"ContentClassifiers": [
"FreeOfPersonallyIdentifiableInformation",
"FreeOfAdultContent"
]
}
},
"OutputConfig": {
"S3OutputPath": "s3://{bucket_name}/{JOB_ID}/output-{manifest_name}/"
},
"HumanTaskConfig": {
"AnnotationConsolidationConfig": {
"AnnotationConsolidationLambdaArn": "arn:aws:lambda:us-east-2:266458841044:function:PRE-TextMultiClass"
},
"PreHumanTaskLambdaArn": "arn:aws:lambda:us-east-2:266458841044:function:PRE-TextMultiClass",
"NumberOfHumanWorkersPerDataObject": 2,
"TaskDescription": "Dear Annotator, please label it according to instructions. Thank you!",
"TaskKeywords": [
"text",
"label"
],
"TaskTimeLimitInSeconds": 600,
"TaskTitle": "Label Text",
"UiConfig": {
"UiTemplateS3Uri": "s3://{bucket_name}/instructions.template"
},
"WorkteamArn": "work team arn"
},
"LabelingJobName": "Label",
"RoleArn": "my role arn",
"LabelAttributeName": "category",
"LabelCategoryConfigS3Uri": ""s3://{bucket_name}/labels.json""
}
I think my Lambda function is wrong, when I change to aws arn (preHuman and Annotation) everything works fine.
This is my afterLabeling Lambda:
import json
import boto3
from urllib.parse import urlparse
def lambda_handler(event, context):
consolidated_labels = []
parsed_url = urlparse(event['payload']['s3Uri']);
s3 = boto3.client('s3')
textFile = s3.get_object(Bucket = parsed_url.netloc, Key = parsed_url.path[1:])
filecont = textFile['Body'].read()
annotations = json.loads(filecont);
for dataset in annotations:
for annotation in dataset['annotations']:
new_annotation = json.loads(annotation['annotationData']['content'])
label = {
'datasetObjectId': dataset['datasetObjectId'],
'consolidatedAnnotation' : {
'content': {
event['labelAttributeName']: {
'workerId': annotation['workerId'],
'result': new_annotation,
'labeledContent': dataset['dataObject']
}
}
}
}
consolidated_labels.append(label)
return consolidated_labels
Are there any reasons?