TAVYNO is built to serve thousands of restaurants at the busiest moment of the week: Friday, 7pm. No stalling widget, no lost reservation, no black screen mid-service.
Status updates live. During an incident you'll read what's happening here within 5 minutes.
Every core service is checked each minute from multiple locations. What you see here is what your floor manager sees.
Reservations & widget
Bookings, waitlist, confirmations
99.99%
App & floor plan
Dashboard, floor plan, guest profiles
99.98%
Messaging (email & WhatsApp)
Confirmations, reminders, newsletter
99.97%
Reporting & analytics
GA4, campaigns, revenue insight
99.95%
Monthly uptime across all core services, plus every incident we logged. We publish the bad months too.
Mar
100%
No incidents
Apr
99.98%
1 incident
May
100%
No incidents
Jun
99.94%
1 incident
Jul
100%
No incidents
Aug
99.99%
1 incident
8 Aug 2026
Minor22 minDelayed WhatsApp confirmations
A provider-side queue at Meta slowed confirmations. Emails went out as normal; all WhatsApp messages were delivered afterwards.
Fix: Automatic fallback to email now triggers after 60 seconds instead of 5 minutes.
14 Jun 2026
Degradation51 minSlow floor plan during Friday service
A database query on large guest histories caused delays of up to 4 seconds in the app. Bookings and the widget kept working normally.
Fix: Query rewritten and index added; load test extended to 10x peak traffic.
3 Apr 2026
Minor18 minPlanned maintenance reporting
Analytics and exports were briefly unavailable during a database upgrade. Announced 7 days in advance, outside service hours.
Fix: Reporting now runs on a read replica, so future upgrades cause no downtime.
No unresolved incidents. Average time to resolution: 34 minutes.
Our commitments are written into the data processing agreement and the service level. No fine print about 'reasonable effort'.
No magic - just choices you make at scale before you need them.
Widget and app are served from edge locations across Europe. Static parts come from cache, so a spike of 500 requests per second never even reaches the database.
Row-level security means each restaurant only ever sees its own data. Connection pooling keeps thousands of concurrent sessions on a handful of database connections.
Heavy work - mailings, exports, network sync - goes to a queue with per-venue limits. One venue sending 40,000 newsletters never slows down your bookings.
Every booking runs through a transactional capacity check with a lock on the time-slot row. Two guests grabbing the last table in the same second: exactly one gets through.
CRM, reporting and analytics read from replicas. Heavy filtering in your guest database never slows down service.
Error rate, queue depth and response times are all alerted on. In practice we see an issue before you notice it.
Restaurants from bistro to starred kitchen run their service on TAVYNO every day. Nine out of ten arrive through a recommendation from another owner.
15 years of hospitality experience combined with a platform that contractually commits to 99.9% uptime. We show you our architecture, not just our slides.
Free migration audit of your current system, guest database moved at no cost and your widget installed - all before your first invoice.
Run a week in parallel with your current system. Not right? Stop at no cost. Monthly terms, no setup fee.
No ticket number, but one named contact who knows how your room runs and simply picks up during service.
We personally guide a maximum of 25 restaurants per month, so every switch goes well. Once full, it's next month.
No commission on your guests, no marketplace reselling them to the place down the street. We only earn when you stay.
The app keeps showing the current service on your tablet and syncs as soon as the connection returns. Online bookings keep coming in and are waiting for you once you're back.
The widget loads under 100 kB and shows availability within a median of 300 ms. It never slows down your own website: it loads independently of your page.
You see the status live here and get notified immediately for a P1. Reservations arriving during an incident are buffered and processed afterwards - no booking is lost.
Yes. We load your existing guest database into a test environment first, you check it, and only then it goes live. You say yes when it's right.
We're happy to spend half an hour with you and your IT or web partner. You get to see the architecture, the service level and our backup strategy - even if you end up staying with your current system.
Book a technical intro