Execution time vs long tasks

[zeman]

It would be super interesting to dig deeper into execution time if possible and highlight any blocking CPU (long tasks > 50ms) caused by a third party. Are long tasks captured by Lighthouse somewhere in the bootup time audit?

We're doing a bunch of third party analysis and just because a third party has a lot of CPU usage doesn't mean it's a bad actor. Only long tasks impact user experience and push out metrics like TTI.

In this example of CPU usage that splits out blocking vs non-blocking CPU, should Drift have the higher impact or Google Tag Manager? Long tasks can dramatically change how execution time is interpreted.

[patrickhulce]

Good points @zeman! I'm thinking an interesting metric could be X ms executing longer than 50 ms.

This would indeed have to land in Lighthouse first 👍