I don’t know anything about this talk, it’s probably excellent. But there is no way a graph like this can ever be a sign of a healthy system, regardless of what kind of subsystems, modules or services are used. Don’t put yourself in a situation like this if you can avoid it.
Replying to @msuriar
Uber microservice Graph. #SREcon
4
25
2
71
Seems like the only voice of reason here imo. Few people operate at the scale Uber do, Netflix have a very similar system graph of their systems.
1
While you’re right that a single bad screenshot of a diagram should not be the basis of an architecture review, scale is not the only measure of complexity. Many boring, traditional companies with a few thousand users have systems that are far more complex than Uber or Netflix.

Mar 27, 2019 · 12:36 PM UTC

1
Replying to @stilkov @Uber
No doubt, my reservation is your tweet implied that Netflix or Uber's systems are unhealthy because of the graph. Ignoring all the context around it (size of the organisation etc). Not to mention "healthy" is a very vague term too.