PerformanceObserver()

callback

A PerformanceObserverCallback callback that will be invoked when observed performance events are recorded. When the callback is invoked, the following parameters are available:

entries

The list of performance observer entries.

observer

The observer object that is receiving the above entries.

droppedEntriesCount Optional

The number of buffered entries which got dropped from the buffer due to the buffer being full. See the buffered flag.

A new PerformanceObserver object which will call the specified callback when observed performance events occur.

The following example creates a PerformanceObserver watching for "mark" (PerformanceMark) and "measure" (PerformanceMeasure) events. The perfObserver callback provides a list (PerformanceObserverEntryList) which allows you get observed performance entries.

function perfObserver(list, observer) {
  list.getEntries().forEach((entry) =>  {
    if (entry.entryType === "mark") {
      console.log(`${entry.name}'s startTime: ${entry.startTime}`);
    };
    if (entry.entryType === "measure") {
      console.log(`${entry.name}'s duration: ${entry.duration}`);
    };
  });
}
const observer = new PerformanceObserver(perfObserver);
observer.observe({ entryTypes: ["measure", "mark"] });

You can use PerformanceObserver with a buffered flag to listen to past performance entries. There is a buffer size limit, though. The performance observer callback contains an optional droppedEntriesCount parameter that informs you about the amount of lost entries due to the buffer storage being full.

function perfObserver(list, observer, droppedEntriesCount) {
  list.getEntries().forEach((entry) => {
    // do something with the entries
  });
  if (droppedEntriesCount > 0) {
    console.warn(`${droppedEntriesCount} entries got dropped due to the buffer being full.`);
  }
}
const observer = new PerformanceObserver(perfObserver);
observer.observe({ type: "resource", buffered: true });

Usually, there are a lot of resource timing entries, and for these entries specifically, you can also set a larger buffer using performance.setResourceTimingBufferSize() and watch for the resourcetimingbufferfull event.