Rancher lb … I recreated mine with global scheduling and it’s properly split across three hosts. Two are fine in my ELB, and one is failing. Or it was yesterday, looks like it’s since gone healthy.
How do people troubleshoot the haproxy setups? I’m more used to nginx and F5s myself.
And what do people do for logs? I would like to be able to centralize logs for a given service stack to see the nginx access and error logs across all 3 hosts. I know rancher supports tracking container logs but I think what I want is one step beyond that. I’m thinking maybe Splunk.com cloud or elasticsearch, if I can leverage AWS’ offering. I don’t want to bloat my images, though. Ideally I could just add syslog support and have a variable for the host, and then people can plug in whatever solution they want.