1. 4

  2. 4

    Ok, I’ll ask. It allows you to ship stuff from docker containers to a UDP socket. Now, syslog has supported sending logs over UDP for quite literally, years now. I assume logstash can speak syslog, and syslog also is great at forwarding stuff, has multiple RFCs, and the whole 9 yards, so what is this buying you?

    Granted, syslog may be more complicated than necessary here, or you might cite syslog’s host of problems, but why is this method better?

    note: I know very little about docker containers, so it’s possible that there’s some reason that the above isn’t viable.

    1. 3

      I’m left with the same questions. I feel like maybe this is part of the ‘Go is the node.js of 2014-2015’. Where everyone is super excited about an entirely mediocre[1] language, and is reinventing every wheel available in it. It’s the same thing that happens with most new languages/tools, I think. Give someone a new toy hammer, and everything looks like a brightly colored nail.

      [1] not that being mediocre is bad, mediocre is good, and Go has some nice features (I really like the idea of go fmt, and goroutines are kind of neat too).

      1. 1

        Couldn’t agree more. That said, my day job is working on logging (and metrics) infrastructure where syslog isn’t always the solution[0]. I just don’t see why syslog isn’t the solution for this specific task. :)

        [0] we even do funny things with syslog like forward whole facilities over log-shuttle to logplex, to name a few.