The world at large seems sold on microservices as a way to build large, quality
systems at the speed needed to compete in today's market. However, microservices
are not without their downsides, one of which is the difficulty of reasoning about and optimising the performance of mircoservices working in tandem.
In this session, attendees will learn the key concepts needed to measure the
performance of their services, identify potential bottlenecks and take
corrective action to ensure services perform as needed.
Before we can reason about the performance of connected services, we must
understand how to reason about services in isolation. The first half of the
session is dedicated to discussing how to measure service performance correctly,
what metrics matter and how to compare the performance of a service over time.
In the second half of the session, we take our single-service knowledge and
expand it out to reason about a set of services working together to provide some
aggregate function. We'll see how the performance of each service affects the
performance of the whole, and learn how to identify where to focus our
optimisation efforts.
|