Kudos to Stefan for having the wisdom and energy to go down that rabbit hole.
I recently added an APM screen to a product to show certain execution timings in a convenient manner.
But the primary timing of interest, page load time, turned out to be fine, and it was auth time across each page that was a problem (it's twice as fast now), similar to what Stefan found.
So exposing performance data is usually useful, despite the initial and intended reason behind doing so.
I recently added an APM screen to a product to show certain execution timings in a convenient manner.
But the primary timing of interest, page load time, turned out to be fine, and it was auth time across each page that was a problem (it's twice as fast now), similar to what Stefan found.
So exposing performance data is usually useful, despite the initial and intended reason behind doing so.