,
8 tweets,
3 min read
Read on Twitter
#chaosday19 : More CapitalOne conversation associated with Single Sign On services. We're hearing about chaos journey for this service.
#chaosday19 : A few elements to cover: operating point, MTTR, and human ops
#chaosday19 : Know your operating point (KYOP) is critical for high reliability organization. Based on Rasmussen Dynamic System model as well as Golden Signals (latency, rate, errors)
#chaosday19 : Focus on MTTR - goal to recover automatically gracefully and quickly. Focused on increase transactional resiliency with circuit breakers, increase yield - drop what you can't support (reduce hard dependencies, simplify capacity), high fidelity failover
#chaosday19 : To improve human operations there's failure obsession with a resiliency council and weekly tabletop drills [ed: the concept of weekly drills is awesome in my mind. Nothing builds expertise like actually practicing]
#chaosday19 : LEarning from experimentation - read and connect timeouts, tune hostname lookup. Failover and failback - automated fast failover and manual controlled fallbacks. Good opportunity to tune alerts.
#chaosday19 : When lightning occurs you see the flash before you hear about it. Just like clients feeling the pain prior to alerts. Chaos engineering can help beat the speed of incident impacts!
@threadreaderapp unroll thank you!
