This week I spent so much time trying to get lambda to work with dynmodb and elasticache. Everything worked well with lambda and dynamodb until I tried to integrate elasticache.

While I was working with integrating elasticache and lambda, I created a VPC and put lambda and elasticache in the same VPC. I didn’t realize that’s when everything broke. Looking back, it feels so stupid. I was trying to debug why I couldn’t connect to redis.

    redis_name = os.environ.get('REDIS_NAME')
    logger.debug('table name: {0}'.format(table_name))
    logger.debug('redis name: {0}'.format(redis_name))
    dynamo_client = boto3.resource('dynamodb', region_name=region)
    rconn = redis.Redis(host=redis_name, port=6379, db=0)
    dynamo_table = dynamo_client.Table(table_name)

I can see redis name getting logged into cloudwatch logs but then lambda timeout. Finally, I used my brain and decided to comment redis.Redis from the code. Then it hit me like a ton of bricks, it wasn’t redis connection timeout but dynamodb.

After that debugging got easier, because placing lambda into a VPC caused it to loose connection to the internet. In order to fix it, I had to re-create VPC with internetgateway and NAT so that lambda can connect to the internet. Apparently, lambda connects to the dynamodb over the internet. Who knew !