How to comfortably monitor 5 daemons while developing them?

Question

This is the scenario: I'm on this project for which I wrote 4 Node.js scripts that act as "servers" for some service (they have to be separated). I also use Redis, that is another service to keep an eye on. The system is not yet in production (the reliability of it will raise another set of problems, but not now).

While developing, those services spits a lot of interesting console.log() messages that I care of. I'm actually using a big terminal window split in 4 with GNU screen, but every time I have to start screen, ctrl+a+c, ctrl+a+c, ctrl+a+c and so on then manually start the services.

How would you comfortably automate this operation? Or are there some tools to better enhance MY user experience as the shepherd of those little monsters? :) I'm on my way using ANOTHER service that will publish those logs via HTTP. The logs (mostly debug messages) will be published by the other 4 guys in a Redis channel that the log2http server will subscribe to...

Answer 1

All of the actions you take at startup can be scripted in your .screenrc:

screen -t Log-A  7 /program/that/reads/log/a
split
other
screen -t Log-B  8 /program/that/reads/log/b
split
other
...etc...

Consult the Screen User's Manual for a full list of the commands and what they do.

Answer 2

Since you are still investigating what tools to use for monitoring, i would suggest taking a look at Splunk at http://www.splunk.com/product .

My own experiences with it are limited, but it seems to be able to do real-time monitoring, plus it installs on several platforms.

Maybe that can replace some of your infrastructure, but notice that it comes with a 60-days free trial only, after that you will have to start paying for the service.