Kinesis stream consumer channelize through redis along with aws autorefreshable session
- python >= 3.0
- boto3 >= 1.13.5
- kinesis-python >= 0.2.1
- redis >= 3.5.0
Install with:
pip install kinesis-stream-consumerOr, if you're using a development version cloned from this repository:
git clone https://github.com/harshittrivedi78/kinesis-stream-consumer.git
python kinesis-stream-consumer/setup.py installThis will install boto3 >= 1.13.5 and kinesis-python >= 0.2.1 and redis >= 3.5.0
There is two consumer which has to be run parallelly one is kinesis consumer and second is records queue consumer (redis). I have added a example.py file in this code base which can be used to check and test the code.
Here you can scale record queue consumer as per your per second message quantum. But kinesis consumer would be only one under which one process per shard will be running. Every process will be consuming the messages from shard which is assigned to it.
import threading
from kinesis_stream.consumer import KinesisConsumer
from kinesis_stream.record_queue import RecordQueueConsumer
from kinesis_stream.redis_wrapper import get_redis_conn
redis_conn = get_redis_conn(host="localhost", port=6379, db="0")
stream_name = "test-kinesis-stream"
region = "eu-west-1"
kinesis_consumer = KinesisConsumer(stream_name, region, redis_conn)
record_queue_consumer = RecordQueueConsumer(stream_name, redis_conn)
kinesis_consumer_thread = threading.Thread(name='kinesis_consumer', target=kinesis_consumer.start)
kinesis_consumer_thread.start()
record_queue_consumer_thread = threading.Thread(name='record_queue_consumer', target=record_queue_consumer.start)
record_queue_consumer_thread.start()Override handle_message func to do some stuff with the kinesis messages.
from kinesis_stream.record_queue import RecordQueueConsumer as BaseRecordQueueConsumer
class RecordQueueConsumer(BaseRecordQueueConsumer):
def handle_message(self, message):
# your code
print(message)Horizontal scaling for kinesis consumer.