I started using "The Cloud" in 2012 to build apps. At the time they called it Google App Engine. In the last 13 years a lot has changed and I've used all the main cloud offerings.
I've been a part of many outages that originated on our cloud provider. And what I've learned is the best practices rarely save you. We pretend multi-az, global DBs and fail overs protect us, but they don't. They make us more scale able, but not more resilient.
These cloud services have great uptime generally, and they have become so centrally integral to the web that when they go down, everything goes with them
If a random person was hosting their own server somewhere, they would probably avoid the huge outages, but likely have more frequent smaller outages and nowhere near the reliable uptime that the big providers have, right?
The Google Cloud post mortem and discussion: https://news.ycombinator.com/item?id=44274563
https://news.ycombinator.com/item?id=44260810
I started using "The Cloud" in 2012 to build apps. At the time they called it Google App Engine. In the last 13 years a lot has changed and I've used all the main cloud offerings.
I've been a part of many outages that originated on our cloud provider. And what I've learned is the best practices rarely save you. We pretend multi-az, global DBs and fail overs protect us, but they don't. They make us more scale able, but not more resilient.
It is difficult to measure
These cloud services have great uptime generally, and they have become so centrally integral to the web that when they go down, everything goes with them
If a random person was hosting their own server somewhere, they would probably avoid the huge outages, but likely have more frequent smaller outages and nowhere near the reliable uptime that the big providers have, right?
[flagged]
Old news.