Keeping Services Running

Pizza, beer, and mingling

"Service Invocation - How to Start and Stop Services"
Evan Miller - Devtools and Deployments Guy, Yahoo!

"Monit Overview"
Rory Toma - Embedded Software Engineer, Ooma

"Using Monitoring and Configuration Management to automatically recover failed services"
Greg Retkowski - Operations Engineer, OnLive

We've had some past talks about scalability and availability, but I've recently heard about some tools for managing availability on a smaller scale. I was searching for solutions to keeping service daemons running on linux, and found three: cron jobs that test and restart failed processes, daemontools/runit, and monit. The idea is that on a single host, you'd like to be able to monitor a service process, and restart it if it dies. Of course, this is no substitute for redundancy in order to handle the case where your hardware melts down, but if your hardware is ok, it'd be nice not to have to get up just to restart something, right?

So, I'm looking for speakers on this topic. If you can speak on runit/daemontools, or monit, that'd be great. Or, if you know of other tools with this functionality, that would be great to hear about too. Contact me through the meetup site if you can put together a 20-25 minute talk to speak on this.