Element reference

Execution context

Execution context

An execution context describes the environment used for testing: browser, operating system, address, product build, and related snapshot axes.

Two layers

Catalogs are shared dimensions: browser, OS, screen resolution, device, User-Agent, locale, deployment target. Some values ship with the service; a project can add its own entries.

An execution context snapshot records what applied at a moment: chosen catalog entries, host, base URL, protocol, and build version. History should not drift if a catalog label is renamed later.

Environment axes

AxisSnapshot fieldExamples
Browserbrowser_context_idChrome, Firefox, Safari
OSos_context_idWindows, macOS, Linux, Android, iOS
Screen resolutionscreen_resolution_context_id1920x1080, 390x844
Devicedevice_context_idDesktop, iPhone 15, Pixel 9
User-Agentuser_agent_context_idDesktop Chrome, iPhone 15
Localelocale_context_idru-RU, en-US
Deployment targetdeployment_target_context_idstaging, production, preview

Relations

Related elementHow it connects
ProjectSnapshot belongs to a project
BugOptional link on the bug Execution context tab
Autotest runRequired when you create a run

Device and User-Agent

These are different axes.

AxisAudienceOn an autotest run
DevicePeople: how to describe the environment on the cardThe worker ignores it. A Playwright phone profile is not turned on from Device alone
User-AgentWorker: HTTP header and mobile / touch signalsApplied. Empty User-Agent means the normal header of the chosen browser engine
Screen resolutionWindow sizeApplied together with User-Agent

Chromium plus User-Agent “iPhone 15” is Chromium with an iPhone header, not Safari on a phone. For a phone-like run, set User-Agent (and screen if you need a specific size). Device may stay empty.

Required fields for autotest runs

Autotest run creation needs an execution context snapshot in the project (with protocol and starting domain). Without a selected snapshot the run does not start.

Individual axes inside the snapshot may be empty. The worker then applies defaults:

Axis in the snapshotIf empty on an autotest run
BrowserChromium
Screen resolutionSize from the selected User-Agent when that User-Agent has a window size; otherwise 1920x1080
User-AgentNormal header of the chosen engine
DeviceUnused by the worker
OS, locale, deployment targetDo not drive the worker browser session

Supported browser keys for the worker (current delivery): chromium, firefox, safari (Playwright WebKit, not Apple Safari on macOS/iOS), webkit, headless_chrome. The create-run Help tab also lists the current set.

Rules

  • Choosing a User-Agent can offer to write screen resolution and Device. Choosing a Device that has exactly one linked User-Agent can offer User-Agent and screen.
  • On an autotest run, Device is ignored by the worker; User-Agent is applied (with screen resolution). Device still describes the environment for people.

Version 1.1.0 · Last updated 2026-09-11

Did this page help?