metric values generated with package differ from those generated with the builtin LightHouse tool #12601
-
|
Howdy! I have a FE application that loads Web Components and I am testing the performance of my project with this tool. I want to report an issue that I have detected when it comes generating timing metrics locally on a Chrome browser installed in your local machine: the values generated locally with this npm package are different from those you get using the built-in chrome browser tool. This is particular to first-meaningful-paint, first-contentful-paint, largest-contentful-paint, and speed-index. Is this its normal behavior, are we meant to throttle up / down anything? Thank you in advance for your help and comments. |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 2 replies
-
|
I have uploaded a video clip showing the differences https://youtu.be/rwUN62gVitg thank you again |
Beta Was this translation helpful? Give feedback.
-
|
Those runs have different settings. Comparing desktop v. mobile is comparing two different things. See the CLI options docs for how to run desktop (hint: look at the You might also be interested in reading the variability documentation about comparing different runs. |
Beta Was this translation helpful? Give feedback.
-
|
that was my mistake, I had taken for granted that desktop was set as the default value for property |
Beta Was this translation helpful? Give feedback.
-
|
thank you! |
Beta Was this translation helpful? Give feedback.
Those runs have different settings. Comparing desktop v. mobile is comparing two different things. See the CLI options docs for how to run desktop (hint: look at the
--preset=desktopoption).You might also be interested in reading the variability documentation about comparing different runs.