#chaosday19 Moving on to applying some chaos to Kube clusters! 
#chaosday19 Glad to see the lightning talks are getting into some direct implementations over theory. Don't get me wrong I love theory but I really really like seeing the 'art of the possible' in action.
#chaosday19 Kube Monkey works by building a schedule of terminations based on opt-in status, chaos paramters and blacklist. Basically what to terminate and when. Terms are scheduled randomly across a configurable time window. Thankfully it does a sanity check right before term
#chaosday19 FAirly simple deployment for Kube-monkey, it's stateless. Config is a single TOML file with about 6 aprameters (dry_run, run hour, start hour, end hour, blacklist, and timezone)
#chaosday19 [ed: would really love a quick comparison of kube-monkey to other chaos tooling that's available on the market for Kube out in the OSS world]
#chaosday19 service owners can opt-in by adding labels (termination mode and blast radius) - doesn't require further configuration of kube-monkey
#chaosday19 kube-monkey continues to be enhanced including PDB aware attacks, opt-out mode vs. opt-in, and add additional attacks.
#chaosday19 you can find more info at asobti/kube-monkey on github or kube-monkey channel on the k8s Slack /fin
• • •
Missing some Tweet in this thread? You can try to
force a refresh
