Cursor's WAL-on-S3 is all the rage, so it's a fun coincidence to see Cloudflare Durable Objects doing the same for years.
Key takeaway is how they reduce durable latency by writing to followers that upload to S3 if primary fails. h/t @KentonVarda
I'm pretty sure the only reason to use Kafka for telemetry data in 2026 is muscle memory.
A simple, open source, S3-backed pipeline can shuttle 1Gb/s of logs to @ClickHouseDB for $200/mo with 2.8s p50 latency. Fast enough for us.