,
7 tweets,
2 min read
Read on Twitter
Yes. We talk a lot about how *necessary* observability is for large distributed systems. We don't talk as much about how much faster and more decisively you can move from day one. How much better your decisions are.
The reason I knew I had to start honeycomb wasn't because of the hair on fire, site-is-down use cases. Even tho that's often the sales pitch for observability.
It was because i literally cannot imagine going back to engineering without it.
How do you decide what to build? How do you know if it's working as intended or not? With dashboards?? LMFAO.. and spend half my day hunting down answers to basic questions? Thanks but no.
How do you decide what to build? How do you know if it's working as intended or not? With dashboards?? LMFAO.. and spend half my day hunting down answers to basic questions? Thanks but no.
People who haven't had event-level observability have no idea how much detail and confidence they are missing out on. How much they depend on intuition and flawed mental models to navigate their systems.
Once you have seen all the shit your time series aggregates are hiding from you, you'll never trust them again.
I would rather drive without my glasses than build without observability. (And I am blind as fuuuuuck)
I would rather drive without my glasses than build without observability. (And I am blind as fuuuuuck)
Relevant sticker #1: 
Relevant sticker (and Nietzsche quote) #2:
