Launching today
Jyotish Meet
High-concurrency booking platform with zero double-bookings.
11 followers
High-concurrency booking platform with zero double-bookings.
11 followers
A production-ready booking platform built with Next.js & Firebase. Featuring a strict Redis atomic locking layer to eliminate race conditions. Open for community testing and real-time performance feedback!
How does the Redis locking layer handle cases where a user’s session drops mid-booking? Genuinely curious if that can lead to stuck slots or if there’s a built-in TTL fallback.
@mervekatn2mai Great question! It handles it in two ways:
Built-in TTL: Every Redis lock has a strict 15-minute expiration (EX 900). If a session completely drops, the slot automatically frees itself after 15 minutes.
Instant Release: If the payment fails or the user manually closes the checkout modal, the frontend catches it and instantly calls redis.del(lockKey) to unlock the slot immediately.
Bottom line: Worst case is a 15-minute hold, but slots never get permanently stuck!
Curious how the Redis locking actually holds up when two users hit the same time slot at the exact same millisecond — would love to know what happens on the edge cases.
Really cool to see a booking platform with proper Redis locking baked in, that's a detail most skip and end up regretting. One thing I'd love to try is calendar sync with Google or Outlook, so bookings can show up alongside existing appointments without manual copying. Would make this feel way more plug and play for consultants already living in their calendars.
@snbilek8382 Thanks so much! Yeah, building the Redis lock took some work, but it was 100% worth it to prevent double-bookings.
You read my mind on the calendar sync! Adding Google and Outlook integration is the very next feature I am building for V2 so it's fully plug-and-play. Thanks for checking it out!
Booked a slot during a busy hour and it locked in cleanly with no double booking, which is honestly more reliable than most paid schedulers I've tried. Nice job shipping the Redis piece without it feeling half-baked.