Change Details

We are seeing strong signals both in RUM data collected from visitors and our synthetic tests that 1.36.0-wmf.10 caused a major performance regression during the initial page load on the mobile site. This seems to manifest itself in the form of early long tasks that freeze the browser multiple times (including downloads) before visual completion. RUM data: {F32363116} Synthetic data: {F32363118} {F32363120}

We are seeing strong signals both in RUM data collected from visitors and our synthetic tests that 1.36.0-wmf.10 caused a major performance regression during the initial page load on the mobile site. An extra 150ms for p75 loadEventEnd on RUM data. Much worse impact on our synthetic tests that simulated slow internet connections. This seems to manifest itself in the form of early long tasks that freeze the browser multiple times (including downloads) before visual completion. RUM data: {F32363116} Synthetic data: {F32363118} {F32363120}

We are seeing strong signals both in RUM data collected from visitors and our synthetic tests that 1.36.0-wmf.10 caused a major performance regression during the initial page load on the mobile site. An extra 150ms for p75 loadEventEnd on RUM data. Much worse impact on our synthetic tests that simulated slow internet connections. This seems to manifest itself in the form of early long tasks that freeze the browser multiple times (including downloads) before visual completion. RUM data: {F32363116} Synthetic data: {F32363118} {F32363120}