Maturing Your Data Architecture in a Week: How Bluesky Survived#
SREcon25 Americas digest Program Adjacent pick 25–27 March 2025
Jaz Volpert, Bluesky PBC — Adjacent
- Why: A real account of a small team (~6 engineers) rapidly scaling physical infrastructure under a genuine viral-growth event (1,200% DAU increase in a week) — notable specifically because Bluesky runs its own physical infrastructure rather than purely cloud-native services.
- What: How Bluesky's backend team matured its data architecture in 11 days of intense effort to handle a surge to 1M+ new users/day and 1,600+ events/sec, while running core infrastructure on owned physical hardware. (from abstract only)
- How: Not detailed mechanically in the abstract beyond "rapidly matured data architecture." (from abstract only — thin on mechanism)
- Where applicable: Physical/self-hosted infrastructure under sudden extreme load; a useful contrast case to cloud-elastic scaling stories since Bluesky couldn't simply autoscale.
