Truisms


Truisms
  • What looks like overkill on metrics, observability, tracking, and auditability usually turns out to be valuable in surprising ways. You want to learn even the "unknown unknowns".
  • Granular modularity leads to the ability to build unpredictably awesome patterns. This is true in many contexts and levels.
  • Always try to be conservative in what we generate, and generous in what we accept (classic protocol design).
  • Murphy's law is right: in production, any points that are not failure-resistant (e.g., redundant) will be the first to fail.
  • Therefore, for a service in production, the lack of a high availability architecture is equivalent to no service at all.
  • Watson's Law of Pinball Repair (with apologies to Occam): "It's always the stupid thing". ( e.g., Plug it in! Replace the fuse!)
  • We as colleagues are all adults. Nothing can move forward well without open communication and mutually earned trust.
  • Moments of discomfort, a down phase in a business cycle, lack of consensus in a decision process: these are exactly the points of maximal leverage, and carry the best potential for strong future outcomes.
  • Finally and most obviously: the hottest trends are usually the shortest-lived.
Our only point of power is the present. Learn from the past; influence the future — now.