The Fault In Our Clouds
Fallthrough
How the DNS Planner/Executor Race Led to Data Loss
Kris explains the planner, executors, and garbage collector interplay that allowed an old plan to be reapplied then wiped, deleting DNS records.
First it was GCP in June. Then it was AWS in October. Then it was Azure a week later. It seems that our cloud providers are having outages far more often, and for far longer, than any of us would like. In this episode, Kris, Ian, and Matthew discuss the two most recent outages along with some of their thoughts on the current state of the industry and the future of software.
We continue this discussion in this week's episode of Break! The panel talks about whether seeking a career with a FAANG company is worth it anymore, why building software for your local community is important, and their frustrations with point of sale systems. Watch it on YouTube or listen with your favorite podcasting app! Learn more by going to https://break.show.
EXTRA! EXTRA! There's lots of bonus content in this episode! And if you're a supporter you're getting all of it. In this week's extra chapters the panel talks about whether we all need to be on large cloud providers, frustrations with food delivery app PINs, whether timeouts and retries should be our go to, and why it feels like software is constantly getting worse. Not a supporter yet? Fix that today by heading over to https://fallthrough.fm/subscribe where you'll get not only extra content but also higher quality audio. Sign up today!
Thanks for tuning in and happy listening!
Show Notes:
- AWS Outage Summary: https://aws.amazon.com/message/101925/
- Azure Outage Summary: https://azure.status.microsoft/en-us/status/history/
Table of Contents:
- Prologue (00:00:00)
- Chapter 1: The AWS Outage (00:03:03)
- Chapter 2: Overdependence on Timeouts and Retries [Extended] (00:27:15)
- Chapter 3: Food Delivery app PINs should be Local First [Extended] (00:27:41)
- Chapter 4: The Azure Outage (00:28:11)
- Chapter 5: Do We Actually Need All These Cloud Services? [Extended] (00:39:37)
- Chapter 6: We Are Trapped By Our Own Path Dependence [Extended] (00:40:07)
- Chapter 7: What Is Popular Is Not Necessarily What Is Good (00:40:54)
- Appendix UNPOP: Unpopular Opinions and Panic & Recover (00:42:42)
- Epilogue (01:02:34)
Hosts
- Kris Brandow - Host
- Ian Wester-Lopshire - Host
- Matthew Sanabria - Host
Socials:
- (00:00) - Prologue
- (03:03) - Chapter 1: The AWS Outage
- (27:15) - Chapter 2: Overdependence on Timeouts and Retries [Extended]
- (27:41) - Chapter 3: Food Delivery app PINs should be Local First [Extended]
- (28:11) - Chapter 4: The Azure Outage
- (39:37) - Chapter 5: Do We Actually Need All These Cloud Services? [Extended]
- (40:07) - Chapter 6: We Are Trapped By Our Own Path Dependence [Extended]
- (40:54) - Chapter 7: What Is Popular Is Not Necessarily What Is Good
- (42:42) - Appendix UNPOP: Unpopular Opinions and Panic & Recover
- (01:02:34) - Epilogue
The AI-powered Podcast Player
