April 18, 2026
Now serving 90 kBq, hot and fresh
The becquerel as an SI unit for request rate
Website traffic in “Becquerels”? Nerds cheer, purists gasp, and the memes are radioactive
TLDR: A tech post suggests labeling average website traffic in Becquerels (events per second), not Hertz, to reflect messy real-world flow. Commenters split: pragmatists want per-minute or “rips,” purists argue SI rules, and jokesters revel in radioactive memes—because consistent metrics matter more than syllables.
The internet’s newest fight: should we measure website traffic in becquerels—yes, the radioactive kind? A provocative post argues request rates should be counted per second and, for messy real-world traffic, labeled in Becquerel (Bq) instead of Hertz (Hz). Cue the comment section meltdown. Some are loving the chaos, others are clutching standards manuals like pearls.
The practical crowd led by hirsin asks, why not per minute? Their point: lots of APIs run slower than once per second—nobody wants to talk in “millibecquerels.” Meanwhile, avmich wonders if we can still say Hz for averages, while standards sticklers like nofriend swoop in with a TED Talk on how Hz is “1 per second,” but… 1 what exactly? Radians? Revolutions? The pedantry is strong, and it’s glorious.
Then the comedy hits. bjoli confesses their Android unit converter keeps spitting out “becquerel” because users forget to divide by time. The crowd roars. murkt shrugs at the naming drama entirely—“90krps” (90 thousand requests per second) versus “90kBq”? Same breath, fewer syllables, same vibe—like saying “Mongo” instead of “PostgreSQL.” Add a fan-favorite side quest: invent a new unit called “rips” (requests per second) and call it a day.
Verdict? It’s physics nerds vs. web devs, with humor as the referee and dashboards as the battlefield.
Key Points
- •Dashboards can report inconsistent request rates if the time aggregation period is not fixed, varying with zoom or display settings.
- •The article recommends explicitly specifying the period length in metrics queries so all viewers see the same rate.
- •Seconds should be the standard period, expressing request rates as requests per second.
- •Two SI units equal to s−1 are compared: hertz (for periodic events) and becquerel (for stochastic averages).
- •Use Hz for regular load tests and Bq for organic traffic; although Bq is defined for radioactivity, the author advocates its practical use for request rates.