Install

pip install s3logs

Usage

s3logs config.conf

Where config.conf can use that structure:

[S3]
access_key = <S3_KEY>
secret_key = <S3_SECRET_KEY>
host = <s3.example.com>
bucket = <bucket_name>
chunk_size = <bytes, default=52428800>

[logs]

suffix = .0.gz
key_suffix = .gz
directory = /var/log/nginx/
filename = <filename, default=(yesterday as yyyy-mm-dd)>

[map]

example.com-access.log = example/access
example.com-error.log = example/error
mysite.me.access.log = mysite/access

When it used with that config, script takes all files in directory /var/log/nginx/, filter only those, which ends with .0.gz and send it to S3, according to map.

For example, /var/log/nginx now consists of:

example.com-access.log
example.com-access.log.0.gz
example.com-access.log.1.gz
example.com-error.log
example.com-error.log.0.gz
example.com-error.log.1.gz
mysite.me.access.log
mysite.me.access.log.0.gz
mysite.me.error.log
mysite.me.error.log.0.gz

So, if today is 9 December 2015, and your hostname is node1, on your S3 <bucket_name> would be those keys:

node1/example/access/2015-12-09.gz
node1/example/error/2015-12-09.gz
node1/mysite/access/2015-12-09.gz

Because we have not explain how maps mysite.me.error.log.0.gz - it would be skipped.

If there was filename=newfile option in config.conf, keys in S3 would look like node1/example/access/newfile.gz