Summary
- A free allowance can conceal the later cost of egress, retrieval, support, extra capacity and migration.
- Teams should price an ordinary month, a growth month and a full exit before placing operational data in a free service.
The dangerous cloud bill rarely begins with one spectacular charge. It accumulates when a prototype becomes a shared archive, backups multiply, users retrieve more files, or a business needs support that the entry tier does not include. At that point, the application and its data may already depend on one provider’s interfaces and operating assumptions.
The relevant comparison is therefore not “free versus paid.” It is the cost of keeping the workload usable over its life. That includes storage after the allowance, requests and data transfer, recovery tests, access controls, staff time and the labour needed to move elsewhere. Privacy and locality requirements can narrow the available exit routes further.
Before adoption, an owner should record the free-tier expiry conditions, set billing alerts and estimate three scenarios: normal use, a demand spike and a complete export. The export should be tested while the data set is still small. A service that cannot produce a timely, usable copy of the data does not offer a credible low-cost exit.
The next useful evidence is a monthly unit-cost record tied to real usage, not the provider’s headline allowance. Free cloud capacity can be a sensible starting point. Without an exit budget, however, it is also an unpriced commitment.
