sqs-s3-logger

A library to persist messages on S3 using serverless architecture. It is mainly targeted at cheaply archiving low-volume, sporadic events from applications without a need to spin additional infrastructure.

What it’s not

Not a replacement for general logging systems or libraries. Provides no filtering or aggregation.

AWS Alternatives

• Cloudwach Logs
• Kinesis Firehose

Usage

Configure boto3’s credentials as per: http://boto3.readthedocs.io/en/latest/guide/quickstart.html#configuration

Make sure you setup:

• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY
• AWS_DEFAULT_REGION (optionally)

Take a look at main.py.

For help: python3 main.py -h

For example (backup at midnight each Saturday from app-logs queue to app-logs-archive bucket):

sqs-s3-logger create -b app-logs-archive -q app-logs -f app-logs-backup -s 'cron(0 0 ? * SAT *)'

Sending messages to a queue

Ideally you should use another AWS IAM user with permissions restricted to getting SQS queues and writing messages.

import boto3
sqs = boto3.resource('sqs')
queue = sqs.get_queue_by_name(QueueName='<QUEUE_NAME>')
queue.send_message(MessageBody='<MSG_BODY_NOT_EXCEEDING_256KB>')

Limitations

• Maximum SQS message size is limited to 256 KB
• There could be no more than 120,000 messages in a queue at a time.
• SQS messages cannot persist for longer than 14 days.
• Lambda environment has up to 512MB of ephemeral disk capacity.
• By default it does not guarantee correct time-based ordering

You may need to adjust your CRON settings depending on your volume.

Testing

python3 setup.py test

These will use your AWS account to instantiate a temporary integration environment.