Hi Simon! On Wed, Nov 12, 2014 at 05:11:27PM +0900, Simon Horman wrote: > I was aware you had done some work in this area but I wasn't entirely > sure what case you were trying to fix. Thanks for filling in that gap > in my knowledge.
Yep. The agent-check is a really nice feature. Initially we naively thought it could be a simple fork of the health checks but we both learned it was something quite different and serving a complementary purpose. I would not be surprized if in the long term it sits completely separate from the health checks, supports all the extra features we currently cannot configure (such as address/ssl) and that we rework the server status not to depend on hacked combinations of both health and agent but instead try to design something more event-based using on a list of force-up and force-down results. That's still unclear in my head but I'm seeing some room for improving what we currently have. BTW, if someone is interested in implementing a simple agent daemon for linux/unix which would only check files presence, it would be nice, as users are starting to ask how to use it without xinetd/nc/socat, especially when they are not allowed to deploy system-wide changes for a single app. Cheers, Willy
