Practically all modern software is operating as a distributed system, even if it might not look like it at first glance. API calls and TCP/IP requests obviously cross process boundaries and use network connections, but so does accessing storage, policy enforcement, or logging within a cloud-based context...and all those operations can fail, often in interesting ways.

In this track, we'll look at how to tackle these distributed system problems. How do we connect components without adding extra points of failure or impacting scalability? What about failure? If a system grinds to a halt, we'll know - but what about partial/grey failures or intermittent performance degradation? We'll look at using observability to keep on top of our systems while, ideally, not breaking the bank in logging costs.