Chromium v148 Common Issues
As part of our ongoing efforts to improve customer transaction tests after the recent Chromium 148 release, we have identified common issues impacting customer transaction tests, as well as known solution paths. This list will be updated as new solutions emerge.
Only common issues with an identified solution path are included on this page; not all issues are included.
Please contact ThousandEyes Support if you find an issue not listed here.
Custom Header Handling
Tests that rely on custom headers may fail or behave differently because Chromium v148 changes how custom headers are applied during requests. If a test fails, try running it with custom headers removed. If it passes, you can update your test configuration to remove or adjust those headers in order to reduce the likelihood of further failures.
Element Timing and Timeouts
Tests may time out or fail with element-not-found errors because Chromium v148 can render elements slower than your script expects. To fix this, increase the explicit and implicit wait timeouts for the affected steps so the script waits for the page to fully render.
Leaked Password Browser Popup
Automated logins may stall or fail if Chromium v148 shows a leaked-password popup for your test credentials, which blocks interaction with the page. To resolve this, use test credentials that do not trigger the warning if possible.
Window Maximize via ChromeOptions
Window layouts may differ after maximizing in Chromium v148, which can cause elements to be hidden or unclickable. If your script uses driver.manage().window().maximize(), try replacing or supplementing it with the ChromeOptions flag --start-maximized.
Page Load Strategy (Normal vs Eager)
Marker durations may increase significantly because ChromeDriver v148 holds WebDriver commands until the page is fully loaded when using pageLoadStrategy: normal. To fix this, change the strategy to pageLoadStrategy: eager so commands proceed as soon as the DOM is interactive.
Not Attached to an Active Page Error
Tests may sporadically fail with Not attached to an active page errors during actions like clicks or screenshots, especially after a page refresh or logout. This happens because Chrome's new architecture can execute commands before the next page has fully loaded. To fix this, add an explicit wait for the specific element right before you interact with it. Alternatively, you can add a short delay, wait for the document to be complete, or configure the script to ignore screenshot errors.
Zstd Content Encoding Error
Requests may hang or fail to return a usable response because Chromium v148 advertises zstd compression, but intermediate network hops may not decode it correctly. To fix this, add the Chrome flag --disable-features=ZstdContentEncoding to fall back to more reliable encodings.
Last updated