Wide logging: Stripe's canonical log line pattern

Stripe’s pitch is simple: instead of scattering details across dozens of tiny notes, write one big, structured log per request with everything—route, user, feature flags, deploy ID, the works. They say it’s not just for fixing crashes; it powers dashboards and long-term analysis, too. Fans call it a “pre-joined” lifesaver. Critics call it… well, something else.

The thread lit up with big feelings. One camp is shouting that this isn’t logging at all—it’s APM (application performance monitoring) or tracing with a new coat of paint. “It’s not what I would traditionally think of as ‘logging’,” fumes one commenter, arguing that logs tell the story during a request, not just at the end. Another crew rolled in with receipts: Facebook’s Scuba did this ages ago and it was “pretty sick,” says one fan, while an Amazonian notes CloudWatch metric events already look like this. Translation: some see Stripe as pioneering; others see them as late to the party—just louder.

Still, there’s consensus on one spicy point: logging is undercooked across the industry, and a tidy “summary line” could save engineers from late-night detective work. Meanwhile, the nerdy drama gets delicious: one user asked if the all-important error_slug needs to be short (what if it’s an epic ffmpeg meltdown?), and another wondered if a “one-liner” can represent many tasks in a job. The memes wrote themselves: “One log to rule them all,” “pre-joined like a rotisserie chicken,” and “finally, a log big enough to have its own LinkedIn.”