Testsigma Cloud Testing: Infrastructure, Devices, and Parallel Execution
Testsigma runs tests on its own cloud infrastructure — you don't manage browsers, devices, or execution servers. This guide covers what that infrastructure provides, how to configure it, and what to watch for in production use.
What the Cloud Provides
When you trigger a Testsigma test run, the platform:
- Provisions execution machines (virtual machines or device slots) based on your test plan configuration
- Starts the appropriate driver (ChromeDriver, GeckoDriver, Appium server)
- Executes your test steps, capturing screenshots and logs at each step
- Terminates the session and stores results
You interact with results in the Testsigma dashboard. No setup, no teardown, no infrastructure cost beyond your Testsigma plan.
Browser and OS Coverage
Cloud execution supports these combinations:
Desktop Web:
| OS | Browsers |
|---|---|
| Windows 10 | Chrome, Firefox, Edge |
| Windows 11 | Chrome, Firefox, Edge |
| macOS Monterey | Chrome, Firefox, Safari |
| macOS Ventura | Chrome, Firefox, Safari |
Mobile Web (Emulators):
- Android: Chrome on emulated Android 10, 11, 12, 13
- iOS: Safari on emulated iPhone 12, 13, 14, 15 (simulator)
Mobile Native (Real Devices): Testsigma maintains a real device lab for iOS and Android native app testing. Coverage varies by plan — Enterprise plans include broader device access.
Parallel Execution
Parallel execution runs multiple test cases or test suites simultaneously across different machines or devices. This is the main speed lever for large test suites.
How It Works
When you enable parallel execution in a test plan, Testsigma:
- Distributes test cases across available execution slots
- Runs them concurrently (up to your plan's concurrency limit)
- Aggregates results in a single report
A suite of 200 test cases that takes 2 hours sequentially can finish in 15 minutes with 8 parallel slots.
Configuring Parallelism
In the Test Plan configuration:
- Max parallel machines: How many machines to allocate simultaneously
- Test distribution: By test case, by test suite, or custom grouping
Start with max parallel = number of browsers × 2 as a baseline and adjust based on actual run times.
Session Limits
Your plan determines how many concurrent sessions you can run. Hitting the limit causes excess tests to queue — they don't fail, but they don't run until a slot opens. Monitor your average concurrency against your plan limit to avoid unexpected queuing.
Local Testing with Testsigma Tunnel
Your app may not be publicly accessible — staging environments, localhost, internal services behind a firewall. Testsigma's local testing tunnel lets cloud machines reach these environments.
Setup
- Download the Testsigma local agent for your OS (Windows, macOS, Linux)
- Start the agent with your account credentials:
./TestsigmaAgent \
--jwtApiKey "$TESTSIGMA_API_KEY" \
--tsAgentKey "$TESTSIGMA_AGENT_KEY"- In your test plan, set Execution Type to Hybrid and select your local agent
Cloud machines route traffic through the agent to reach your internal URLs.
Tunneling Limitations
- The tunnel adds latency (typically 50-200ms per step depending on network)
- Each agent handles one test session at a time by default; parallel execution on local agents requires multiple agent instances
- Firewalls must allow outbound connections to Testsigma's relay infrastructure
Test Environments
Test environments in Testsigma store configuration that varies between deployments: base URLs, API endpoints, credentials, feature flags.
Create environments for:
- dev —
https://dev.yourapp.com, dev credentials - staging —
https://staging.yourapp.com, staging credentials - production —
https://yourapp.com, read-only test credentials
Reference environment variables in test steps:
Navigate to ${app_url}/login
Enter ${test_username} in the Email fieldSwitch environments at the test plan level without modifying test cases. Same tests, different targets.
Execution Scheduling
Testsigma supports scheduled test plan execution:
- Cron-based scheduling — run at a fixed time daily, weekly, or on a custom schedule
- CI/CD trigger — run via REST API on every deployment
- Manual trigger — run on-demand from the dashboard
For regression suites, a common pattern:
- Smoke tests: triggered on every deployment (via CI/CD API)
- Full regression: scheduled nightly at 2 AM
- Cross-browser suite: scheduled weekly on weekends
Viewing Execution Results
The execution results dashboard shows:
- Summary: total tests, passed, failed, skipped, execution time
- Step-level details: each step with status, screenshot, and timing
- Video playback (Enterprise): full session recording for debugging
- Logs: browser console logs and Testsigma agent logs
For failing tests, the screenshot at the moment of failure is the fastest debugging starting point. Testsigma captures it automatically.
Downloading Results via API
Pull execution results programmatically:
# Get test plan run results
curl -s \
"https://app.testsigma.com/api/v1/execution_results/123/test_case_results" \
-H "Authorization: Bearer $TESTSIGMA_API_KEY" \
| jq '.content[] | {name: .testCaseName, result: .result}'Export JUnit XML for CI/CD systems:
curl -s \
"https://app.testsigma.com/api/v1/execution_results/123/junit_report" \
-H "Authorization: Bearer $TESTSIGMA_API_KEY" \
-o results.xmlMost CI systems (Jenkins, GitHub Actions, GitLab CI) can parse JUnit XML to display test results natively.
Pricing Considerations for Cloud Testing
Cloud testing costs grow with:
- Number of concurrent sessions (parallelism)
- Run time per session (long tests × many browsers)
- Real device usage (more expensive than emulators)
Estimate your monthly usage: (tests per run) × (browsers) × (avg test duration minutes) × (runs per month). Compare against your plan's included minutes.
For a team running 200-test regression suites across 4 browsers twice daily:
- Sessions: 200 × 4 = 800 per run
- Minutes: 800 × 2 min avg = 1,600 per run
- Monthly: 1,600 × 60 runs = 96,000 minutes
This is well into Business or Enterprise territory. Run the math before committing.
Summary
Testsigma's cloud removes infrastructure complexity from test execution. You get browsers, devices, parallel slots, scheduling, and result storage without managing servers. The tradeoff is cost at scale and less control over the execution environment. For teams that want to run tests without an ops burden, it's a solid infrastructure option. For teams running hundreds of thousands of test minutes monthly, dedicated cloud testing services or self-hosted Selenium Grid may be more cost-effective.