Testing distributed systems

Verifying the correctness of any large system is a non-trivial exercise. Now break that system into individually deployable, interdependent pieces, and try to test all that the usual way.. It is not pleasant and it hardly works.

This talk will cover classical testing, its typical failure modes and in particular its weaknesses when applied to distributed systems. It will then focus on approaching system verification via a multitude of alternative monitoring-based tools, collectively known as testing in production, together with their pros and cons