Native support for external ad/tracker blocklists (URLs/Files) to measure third-party impact

[deffcolony]

Feature/improvement

Is your feature request related to a problem? Please describe.

Currently, if we want to measure the performance impact of third-party ads and trackers on a website, we have to manually specify domains using --browsertime.block or --browsertime.chrome.blockDomainsExcept.

While this works for a handful of domains, it becomes difficult to manage when trying to use large, community-maintained blocklists (like those used by Pi-hole, uBlock Origin, or AdGuard) which contain tens of thousands of domains. Passing thousands of domains manually via CLI arguments is impractical and can hit maximum command-line length limits in some environments.

while Chrome's SPOF feature (--host-resolver-rules) or a custom wrapper script can technically achieve this, there is no built-in, frictionless way to automate this within the tool itself.

Describe the solution you'd like

I propose adding native support for fetching and parsing external blocklists directly within Browsertime/sitespeed.io.

Ideally, this would introduce two new CLI flags:

1. Pass blocklist URLs directly:

--browsertime.blockListUrl https://raw.githubusercontent.com/StevenBlack/hosts/master/hosts
--browsertime.blockListUrl https://raw.githubusercontent.com/PolishFiltersTeam/KADhosts/master/KADhosts.txt

2. Pass a file containing a list of URLs:

--browsertime.blockListFile blocklists.txt

(Where blocklists.txt contains a plain-text list of raw URLs to fetch).

Expected Behavior:

1. Sitespeed fetches the provided URLs before the test starts.
2. It parses the content (supporting standard hosts file formats or basic line-by-line domain lists, stripping comments #).
3. It converts them internally into the equivalent of --browsertime.block rules.
4. The browser drops these requests at the network level during the test.

Describe alternatives you've considered

1. Custom Wrapper Scripts: Writing a custom Node/Bash script that downloads the lists, parses them, formats them into a massive string of --browsertime.block domain1 --browsertime.block domain2, and executes the Docker container. (Drawback: Highly prone to hitting CLI character limits and requires everyone to reinvent the wheel).
2. SPOF Testing via Chrome args: Using --chrome.args='--host-resolver-rules=...' to map trackers to unreachable.sitespeed.io. (Drawback: Browser specific, tedious to format for 50,000+ domains, and slightly different behavior than outright blocking the request).

Additional context

The primary use-case for this is Automated A/B Performance Testing.
By easily toggling these blocklists on and off, developers and performance engineers can easily quantify exactly how much third-party scripts are hurting their site's UX.

It would allow us to easily generate reports comparing:

• Baseline Page Load
• Page Load with trackers blocked

And cleanly observe the delta in metrics like:

• SpeedIndex / LCP
• requests.total
• thirdParty.requests
• transferSize.thirdParty
• cpu.longTasks

Bringing Pi-hole/uBlock style blocking natively to sitespeed.io would be an very useful asset for the web performance community.