Hard Limit

A hard limit works like a debit card that gets declined the instant your account balance hits zero. If you set a monthly spending cap in advance, the moment usage reaches that amount, requests simply stop being processed. Unlike a soft limit, which just sends a warning while letting you keep going, a hard limit actually locks the door.

Why is this kind of safeguard needed? When a company hands out API (the channel programs use to exchange data with each other) keys across multiple apps or teams, a test server might accidentally keep firing requests, or a bug could cause calls to spiral out of control, and the bill can balloon overnight. Since no one can watch the logs around the clock, the system is built to stop itself automatically once it crosses the limit.

That's why hard limits serve as the last line of defense for protecting a budget. They can be set at the organization level, per project, or even per individual key, which prevents one team's mistake from eating into another team's budget.

In the article, it appears in a line like "it also supports a hard limit that blocks traffic itself once the threshold is reached." What's easy to misunderstand here is that this isn't just a warning notification. Once you cross the limit, it's not just an email you get — API requests are actually rejected from that point on.

If you're running a service that uses an API, go to your account management screen and set monthly spending caps separately at the organization level and the project level. It also helps to set a soft limit that triggers an alert at around 80-90% of the cap, so you can avoid the sudden shutdown that comes with hitting the hard limit unprepared.