Test coordinate parallel test execution across multiple environments and frameworks.
Use when performing specialized testing.
Trigger with phrases like "orchestrate tests", "run parallel tests", or "coordinate test execution".
Coordinate parallel test execution across multiple test suites, frameworks, and environments. Manages test splitting, worker allocation, result aggregation, and intelligent retry strategies.
Prerequisites
Test runner with parallel execution support (Jest, Vitest, pytest-xdist, Playwright, or JUnit 5)
CI/CD platform configured (GitHub Actions, GitLab CI, CircleCI, or Jenkins)
Test suite with consistent pass rates (flaky tests identified and tagged)
Sufficient CI runner resources for parallel worker count
Test result reporting tool (JUnit XML, Allure, or equivalent)
Instructions
Analyze the existing test suite using Grep and Glob to catalog all test files, their framework, approximate run time, and dependency requirements.
Classify tests into execution tiers:
Tier 1 (Fast): Unit tests with no I/O -- target under 30 seconds total.
Tier 2 (Medium): Integration tests requiring local services -- target under 3 minutes.
Tier 3 (Slow): E2E and browser tests -- target under 10 minutes.
Configure parallel execution for each tier:
Split unit tests across N workers using jest --shard=i/N or pytest -n auto.
Shard E2E tests by test file using Playwright --shard=i/N or Cypress parallelization.
Assign heavier integration tests to dedicated workers with more resources.
Create a CI pipeline configuration that runs tiers in parallel:
Tier 1 and Tier 2 run concurrently on separate jobs.
Tier 3 runs after a fast pre-check gate passes.
Each tier reports results to a unified collection step.
Implement intelligent retry logic for flaky tests:
Tag known flaky tests with @flaky or equivalent marker.
Retry failed tests up to 2 times before marking as failed.
Track flaky test frequency in a log file for triage.
Aggregate results from all parallel workers into a single report:
Merge JUnit XML files from each shard.
Calculate total pass/fail/skip counts and execution time.
Identify the slowest tests for optimization targets.
Write the orchestration configuration to the project's CI config file and validate it with a dry run.
Output
CI pipeline configuration file (.github/workflows/test.yml, .gitlab-ci.yml, or equivalent)
Test sharding configuration with worker count and split strategy
Merged test result report in JUnit XML or JSON format
Execution timeline showing parallel job durations and bottlenecks
Flaky test inventory with retry counts and failure patterns
Error Handling
Error
Cause
Solution
Shard produces zero tests
Uneven test distribution or incorrect shard index
Verify shard count matches actual test file count; use file-based splitting
Worker out of memory
Too many parallel processes on one runner
Reduce --maxWorkers or -n count; increase runner memory; use --workerIdleMemoryLimit
Test ordering dependency
Tests pass in isolation but fail in specific shard order
Add --randomize flag; fix shared state leaks; enforce test independence
Result aggregation mismatch
Missing shard results due to job timeout
Set job-level timeouts higher than test timeouts; add result upload as a separate step
CI cache miss slowing startup
Dependencies not cached between parallel jobs
Configure dependency caching per lockfile hash; use a shared setup job