Lambda does not trigger on all S3 file uploads

Good day,

I have a lambda function which is configured to trigger for all items created in a specific S3 folder. The event trigger config:

Bucket arn: arn:aws:s3:::dev-signal-bucket
Event types: s3:ObjectCreated:*
isComplexStatement: No
Notification name: has private info
Prefix: my_subfolder/

The first few lines of my lambda function src file handler:

def handler(event, context):
    time.sleep(10)
    try:
        temp_dir = tempfile.gettempdir()
        DOWNLOAD_PATH = os.path.join(temp_dir, 'data.npy')
        bucket = event['Records'][0]['s3']['bucket']['name']
        encoded_key = event['Records'][0]['s3']['object']['key']
        key = urllib.parse.unquote(encoded_key)
        print(f"Handling file: {key}")
        s3.download_file(bucket, key, DOWNLOAD_PATH)
    except Exception as e:
        print(f"Error downloading magnitude spectrum from S3: {e}")

Every 5 minutes, multiple measurements are uploaded to the S3 subfolder my_subfolder from a raspberry pi. Whenever a measurement is uploaded successfully to the S3 subfolder, the device logs the record_timestamp and the device id in a DDB table my-ddb-table. Most of the times, the upload triggers the lambda, which then calculates some values and updates the corresponding item in the DDB table with the calculated values. However, in some cases, the upload is successful, but my lambda function is not triggered. I do not even see the "Hanlding file: ..." in my Cloudwatch logs. I know the upload is successful because the IoT device logs in the DDB table and I can see these files in my S3 subfolder in the AWS console. When I re-upload one of these files manually in the console, the lambda triggers and calculates and stores the values correctly, meaning the files are not corrupted. Also, no errors are logged into my Cloudwatch logs.

My lambda does not throttle.

My question is: why is my lambda not always triggered by a successful S3 item upload?

Help is very much appreciated. If anything is unclear, please let me know. Thanks in advance!

Kind regards, Miranda