Current life in Tokyo: Lots of detail about the event, it’s impact, how they recovered, and how Japan will prepare for the next one. Very well put together.
The cloud under the sea: Story on the Verge about how the undersea fibre optic cable network is built and maintained. I have heard nothing about this ever in my life …
Real Talk: What We Think We Know — That Just Ain’t So: John Allspaw talks about misconceptions in incident response type work.
Why is observability so expensive?: Any source that emits a metric or a log is triggering a process that collects, ingest, stores and analyzes that output. This is a cost that can’t be avoided. Our systems emit a lot of data by design just in the offhand chance that we’ll need it later. Microservices may have exacerbated this problem …
NASA’s Voyager 1 Resumes Sending Engineering Updates to Earth: H/w fault is correct by jpl team on voyager spacecraft. Latency to the craft is 22h 1-way with the round trip being 44hr. That’s one heck of a feedback loop. :)
Investigating the impact of HTTP3 on network latency for search: Dropbox experiments with http3. It helps user that are further away (EU, ASIA) less so for users in NA. They describe their test strategy which was kind of neat. They setup a test domain with a no-op api endpoint that clients would send requests to and then report timings back to Dropbox. They determined network time was a big component in an overall request.

Slides