Can you explain the concept of system availability within the CAP theorem?

In the CAP theorem (Consistency, Availability, Partition tolerance), availability means that every request to the system receives a response — successful or with an error — without guarantees that the data will be current or consistent.

Simply put, the system always responds to requests, even if part of the network is unavailable or a failure has occurred. This is important for services where it is critical not to lose responsiveness, such as in web applications.

Within CAP, during network partitions, the system must choose between:

• Consistency — all nodes see the same data at the same time.
• Availability — the system always responds to requests.

If availability is chosen, then in case of a network partition, the system may serve stale data but will not hang or refuse to respond.