Can you have the cake and eat it too? Fast and cheap observability

With explosive growth of data and micro services, Observability has become expensive, complicated, and slow. This talks explores why it's expensive and has become hard to debug issues quickly. 

This talk explores both the legacy and state of the art approaches to handle cost, and performance in observability systems. It also explains why this is complicated to get it right, why people end up choosing one over the other. Finally, explores with all the recent innovations, if we can achieve both fast and cheap observability. 

We will touch below topics:

- explore various alternatve approaches

- why it's complicated to get both right?

- what were pros and cons of various previous attempts?

- why is it possible to solve the problem now? 

- Are there any innovations and learnings we can borrow from other fields like data warehousing and data streaming?