An endpoint can return 200 OK while the feature it supports is unusable. The
page loads, the health check stays green, and a customer still cannot finish
what they came to do.
That gap is why feature uptime needs a different kind of check.
A response is only one part of an outcome
Endpoint monitoring answers an important question: did a server respond? It can also measure latency, verify a payload, and catch infrastructure failures quickly. Those checks remain useful.
But customers do not experience an endpoint in isolation. They experience a sequence of pages, controls, redirects, and messages. A sign-in journey might depend on a form accepting input, a button submitting it, a session cookie being stored, and the browser arriving at the right account page. Each server in that sequence can be healthy while the journey itself is broken.
The same is true of checkout. An API can accept requests while a disabled button, a looping redirect, or a misplaced validation message prevents every customer from reaching it.
Monitor the promise, not only the machinery
A useful journey check starts with the outcome the product promises:
- a customer can sign in;
- a new account can finish onboarding;
- a shopper can add an item and complete checkout;
- a visitor can find and open the newest article.
Journeys gives that outcome to an agent driving a real Chromium browser. The agent clicks, types, waits, and navigates through the interface, then judges whether the requested result was reached. This checks the feature from the customer’s side of the glass instead of inferring it from one service response.
Evidence should travel with the verdict
Knowing that a journey failed is only the start. A useful alert needs enough context to help someone understand what happened.
A browser run can preserve the activity, screenshots, and recording around the failure. That evidence shows whether the page rendered differently, a control stopped responding, or the journey reached an unexpected destination. It turns “the feature is down” into something a team can investigate.
Use both views of uptime
Feature monitoring does not replace infrastructure monitoring. The two answer different questions.
Endpoint checks watch the services and contracts that make a system run. Journey checks watch whether those parts still combine into an outcome a customer can complete. Keeping both views closes the space where every component looks healthy but the product is not.
For the broader decision, compare the six types of website monitoring and the specific promise each one can support.