Skip to content

Conversation

@Tyriar
Copy link
Member

@Tyriar Tyriar commented Dec 8, 2025

No description provided.

@Tyriar Tyriar added this to the December / January 2026 milestone Dec 8, 2025
@Tyriar Tyriar requested a review from rwoll December 8, 2025 16:17
@Tyriar Tyriar self-assigned this Dec 8, 2025
Copilot AI review requested due to automatic review settings December 8, 2025 16:17
@Tyriar Tyriar enabled auto-merge December 8, 2025 16:18
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR modifies the input latency telemetry collection logic to always report telemetry data when GPU acceleration is enabled, in addition to the existing random 1% sampling. The change treats GPU acceleration as an opt-in experimental feature that warrants additional monitoring.

Key changes:

  • Added a condition to enable telemetry reporting when editor.experimentalGpuAcceleration is set to 'on'
  • Updated the comment to explain this behavior

if (Math.random() <= 0.01) {
// Only log 1% of users selected randomly to reduce the volume of data, always report if GPU
// acceleration is enabled as it's opt-in
if (Math.random() <= 0.01 || this._configurationService.getValue('editor.experimentalGpuAcceleration') === 'on') {
Copy link

Copilot AI Dec 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The GPU acceleration check happens once during construction. If a user enables or disables GPU acceleration after the component is initialized, the telemetry reporting behavior won't update accordingly. Consider either:

  1. Adding a listener for configuration changes to update the behavior dynamically, or
  2. Checking the configuration value in the _logSamples() method each time telemetry is sent (line 83 already reads the current value)

This would ensure that telemetry accurately reflects the current GPU acceleration state rather than the state at initialization.

Copilot uses AI. Check for mistakes.
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Wanted to avoid adding more code, it's fine that this only applies to those that have it on startup only. It'll eventually get removed as more users go on it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants