Frontend reliability is commonly mentioned when it comes to outages. Groups put together for failed API calls, downtime and visual crashes as a result of these failures are simple to acknowledge and measure. Nonetheless, in lots of fashionable purposes, the larger problem is just not full failure however latency. Programs hardly ever go totally offline. As an alternative, they turn into gradual sufficient that customers lose confidence within the interface lengthy earlier than something technically breaks.
Most frontend engineers have skilled this in manufacturing. A web page ultimately hundreds, however solely after a number of seconds of ready. A save motion succeeds within the backend, but the interface stays unchanged lengthy sufficient that the person clicks the button once more. A dashboard renders instantly, however the vital knowledge seems so late that the appliance feels unstable. In follow, customers hardly ever distinguish between “gradual” and “damaged.” If an interplay feels unsure or delayed, belief drops shortly.
As frontend techniques turn into more and more depending on distributed cloud infrastructure, latency turns into a traditional working situation somewhat than an occasional exception. APIs could rely upon a number of downstream companies, serverless techniques could introduce startup delays and state updates could propagate asynchronously throughout areas or caches. Frontend reliability due to this fact can now not be outlined solely by uptime. It additionally will depend on how clearly the interface behaves whereas ready on gradual cloud dependencies.
