I like that it's easy for us as an organization to use with our existing infrastructure without having too many additional dependencies or costs associated (as long you're using Azure). It can be difficult sometimes getting support from them when something goes wrong - but they are always quick at responding! The biggest challenge is just ensuring we have enough resources in different regions so if one fails then there will still remain some availability left within your region of choice. We've been able solve this by provisioning more VM instances across multiple AZs and enabling automatic failover between those zones which has also helped keep down any downtime due to failure events themselves.