For E-band backhaul in a rainy climate, I would start from 99.99% (about 53 minutes of outage a year) as the planning target for a single hop, and 99.999% only where the link carries traffic that cannot be rerouted. The reason is that rain is not a small perturbation at 70/80 GHz: attenuation in heavy rain can exceed 10 dB/km, so chasing five-nines on a long hop usually means either a very short hop or an unrealistic fade margin.
In practice I work from a rain-rate model for the region rather than a single margin number. Take the local rain rate exceeded for 0.01% of the year, convert it to specific attenuation, multiply by the hop length, and then add a margin for wet-antenna loss, which is often 2 to 4 dB per end and is easy to forget. A 2 km hop in a tropical downpour can need 25 to 30 dB of rain fade margin before you even count antenna misalignment or multipath, which is why most deployments keep E-band hops under 2 to 3 km in heavy-rain regions and use a lower band as a fallback path.
Adaptive modulation is what makes the target affordable: design the link so that it drops to a lower modulation and narrower channel during rain instead of dropping out. That turns a hard outage into a throughput reduction, and the availability figure you quote should say which throughput the availability applies to.
What hop length and region are you planning for? With the local rain statistics and the antenna sizes you have in mind, the right margin can be calculated rather than guessed.