Step 5: Launch and monitoring

At launch, Google enables all of your eligible inventory in our production environment. This completes the integration and allows any external user to book or reserve your inventory through the Actions Center.

Once you've launched, it's important to monitor the health of your integration. The following thresholds must be maintained. Failure to maintain these thresholds consistently will result in integration take-down.

Feeds

• Feeds should be sent on a daily basis with no errors or warnings
  • Processing instructions should be set to PROCESS_AS_COMPLETE
  • For Availability feeds, the full inventory daily feed upload should not set any _restrict fields.

Booking server

For all booking server implementations, there is a HealthCheck route that should be included. Google will periodically check your HealthCheck route and should it not respond or return an unhealthy response, we will temporarily disable your integration. We will continue to periodically check your HealthCheck route and once it resumes returning a healthy response we will automatically restore your integration.

Standard implementation
Method	Error Rate Thresholds	Latency Thresholds
CheckAvailability	<10%	<5s
BatchAvailabilityLookup	<3%	<1.5s
CreateLease	<10%	<5s
CreateBooking UpdateBooking	<5%	<4s
CreateBooking (with payments)	<5%	<15s
SetMarketingPreference	<5%	<5s

Real-time updates

For real-time updates, latency is measured by the time difference between when an action is taken (e.g. modifying a booking) and when Reserve with Google receives the real-time update request.

API	Error Rate Thresholds	Latency Thresholds
AvailabilityReplace RTU	<10% each day	<5 mins
BookingNotification RTU	<10% each day & for each state	<5 mins

Error rates can be monitored through the various Partner Portal dashboards, namely the Feeds, Booking Server, and Real-time Updates dashboards.