Blog
Notes on time, measured.
Notes
05
Engineering notes from the Ostrid team: latency, scheduling, and the design of tools that show where time goes.
1 min
Where the 340 milliseconds go
A single agent step, taken apart: network, queue, model, tools. Most of the time is not where people expect.
1 min
Why we stopped quoting average latency
Averages hide the slow runs your users remember. We publish p50, p95 and p99, and here is why.
1 min
A cold start is a product decision
Keeping workers warm costs money. Letting them go cold costs users. How we price the trade-off.
1 min
Designing a console for things that take time
Runs are timelines, not log files. Notes from rebuilding the Ostrid console around a single time axis.
1 min
Notes from moving to two new regions
What changed when Frankfurt and Toronto went live, and the one routing bug we shipped on day two.