How to configure syslog for logging in Python?

System Administration: What limits will be seen when piping Apache logs through logger/syslog for central aggregation?

• I'm expecting to use something like Flume or Spread, but curious as to what limits syslog (syslog-ng, rsyslog, other variants) has with Apache logging.

• Answer:

  I don't know that there's a performance issue with logging via syslog in this context. The real issue is control of data flow, reliability of delivery, and the other features offered by something like Flume. Most syslog implementations are meant to be lean and handle significant transfer, but lack many of the features we'd expect from modern log collection systems. That said, it's simple and it certainly works. For a low dependency, low overhead system for collection of local data (where data loss and routing are major obstacles) syslog works great. For managing tiered collection, high availability, reliable delivery of streaming data, I would opt for Flume. Full disclosure: I'm a Flume committer and work at Cloudera.