It is reliable enough to be our own primary hosting platform for https://www.nexedi.com and to host all our development environments, testing environments and disaster recovery environments.

It is reliable enough if you can deploy your application on two servers, in different DC (eg. Hydro66 and Data4). Rapid.Space CDN and resilient IPv6 backbone can be used for load-balancing.

It is not yet reliable enough to be the primary hosting platform of transactional sites such as https://www.bipandgo.com/. This will change in the future though, once we can have more staff in DCs.

We base our own reliability calculations on a value 8 to 24 hours of unavailability per year. This level is neither high nor low. It is quite similar to the availability of other cloud providers.

Average downtime of all cloud providers is actually higher than what people imagine. In a report conducted in 2012, an average downtime of 9h was measured (https://www.nexedi.com/blog/P-RESILIENCE-NIST.2014). An updated version was published in 2014: http://iwgcr.org/wp-content/uploads/2014/03/downtime-statistics-current-1.3.pdf with similar results.

Even today, downtime remains high in the industry (see https://www.crn.com/slide-shows/cloud/the-10-biggest-cloud-outages-of-2018). 

Rapid.Space downtime is thus comparable to others based on 18-month experience and benchmarking with other cloud providers we also use (AWS, Rackspace, OVHCloud, UCloud, Qingcloud, etc.).