Many of the author's issues with kubernetes are specific comparisons with their current workflow. One I can definitely agree with is the burden of upgrades and

Many of the author's issues with kubernetes are specific comparisons with their current workflow. One I can definitely agree with is the burden of upgrades and keeping the platform current. There are techniques which can make that easier to handle. One statement that surprised me: >> at Google it isn’t uncommon for them to have multi-hour outages with GKE For what it's worth we've been running multiple GKE clusters in production for over three years. We're medium size, with some dozen or so in-house services handling perhaps a total of 20k rps. We are rarely affected by any GCP issue, and as far as I can recall we have never been down due to a GKE-specific problem. In addition to the basic orchestration features we make significant use of ingress and storage primitives. It all quite literally just works.