A message from our sponsor, incident.io:
Ever wonder how Netflix handles incidents at their scale? With incident.io, they’ve built a process that’s smooth, scalable, and keeps everyone on the same page. Tools like Catalog and Workflows make it intuitive for teams to tackle incidents consistently—no matter how big the challenge.
We can never see our systems directly, so we rely on “sensors” to understand the state of the system. What if the sensors are broken?
Lorin Hochstein
Two super insightful observations about the nature of architectural work, well worth revisiting next time you’re making big design decisions.
So, “Two IMO relevant findings regarding architectural work” would probably be a more accurate title. But that would be a lot less catchy title … ;)
Uwe Friedrichsen
To prevent revalidation stampedes, Cloudflare uses randomness to decide whether to send requests to the origin. Click through to find out how it works.
Thibault Meunier — Cloudflare
Some problems with autoscaling, along with potential solutions.
John Akkarakaran Jose — DZone
This article provides a detailed overview of the Incremental Migration with the Dual Write strategy, including the necessary steps, considerations, and best practices.
Deepti Marrivada, Bal Reddy Cherlapally, and Spurthi Jambula — DZone
trying to build the perfect system that anticipates every future need is often worse than creating a system designed to change quickly.
I’ve experienced this firsthand as well. Even an architecture that was supposed to be static needed to change as requirements evolved.
Simen A. W. Olsen — Pulumi
Using more reliable clocks with definite precision allows for significant performance improvements in distributed systems, as described in this article.
Murat
This opinion piece argues that Snapshot Isolation is the “sweet spot” isolation level that is best for most applications.
Marc Brooker