Run E2E tests locally using the new-e2e framework with Pulumi-based infrastructure
62
55%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/run-e2e/SKILL.mdRun E2E tests from test/new-e2e/tests/ using dda inv new-e2e-tests.run.
Parse $ARGUMENTS to determine what to run. The user may provide:
windows/install-test, agent-platform/upgrade, containers)TestInstall, TestUpgrade)Resolve the test target:
test/new-e2e/ to targets, so targets must be relative to that directory (e.g., ./tests/agent-subcommands/flare, NOT ./test/new-e2e/tests/...)agent-subcommands/flare, use ./tests/agent-subcommands/flare as --targetstest/new-e2e/tests/ using Glob, then strip the test/new-e2e/ prefix for the targetTest), find which package contains it using Grep under test/new-e2e/tests/, then set --targets to the package path (relative to test/new-e2e/) and --run to the test nameBuild the command:
dda inv new-e2e-tests.run --targets=./tests/<path> [flags]IMPORTANT: --targets paths are relative to test/new-e2e/. Do NOT include test/new-e2e/ in the target path.
Supported flags (pass through from $ARGUMENTS):
--run <regex> — Only run tests matching this regex--skip <regex> — Skip tests matching this regex--keep-stack — Keep infrastructure up after test (for debugging)--configparams <key=value> — Override Pulumi ConfigMap parameters--agent-image <image:tag> — Use a specific agent image--cluster-agent-image <image:tag> — Use a specific cluster agent image--stack-name-suffix <suffix> — Add suffix to stack name (useful for stuck stacks)--verbose / --no-verbose — Toggle verbose output (default: verbose)--max-retries <n> — Retry failed tests up to n times--flavor <flavor> — Package flavor (e.g., "datadog-agent")--cache — Enable test cache (disabled by default)Before running, confirm the full command with the user.
Run the command with a 60-minute timeout (infrastructure provisioning can take a while). Use run_in_background for the Bash tool since e2e tests are long-running.
After completion, summarize the results: which tests passed, which failed, and any useful error output.
The following must be configured before running e2e tests:
pulumi CLI installed~/.test_infra_config.yaml exists with proper configurationIf any prerequisite is missing, inform the user what needs to be set up.
Tests are organized under test/new-e2e/tests/, running ls test/new-e2e/tests should give the list of test packages
If the test was previously executed and that the infra is in a weird state, that Pulumi is not aware of, trying to rerun the test with the same stack can lead to strange error, like resource being replaced, while they should not exist at all yet. To avoid that issue it is possible to execute the test with a stack with a different name, using --stack-name-suffix <suffix>, please use short stack name suffix. Stop the execution early when you detect that issue in the logs
# Run all tests in a directory
dda inv new-e2e-tests.run --targets=./tests/windows/install-test
# Run a specific test
dda inv new-e2e-tests.run --targets=./tests/agent-platform/tests/upgrade --run TestUpgrade
# Keep stack for debugging
dda inv new-e2e-tests.run --targets=./tests/containers --run TestContainerLinux --keep-stack
# Run with specific agent image
dda inv new-e2e-tests.run --targets=./tests/agent-platform/tests --agent-image "my-registry/agent:latest"
# Run with stack name suffix
dda inv new-e2e-tests.run --targets=./tests/windows/install-test --stack-name-suffix 2
## Usage
- `/run-e2e windows/install-test` — Run all Windows install tests
- `/run-e2e windows/install-test --run TestInstall` — Run only TestInstall
- `/run-e2e TestUpgrade` — Auto-find and run TestUpgrade
- `/run-e2e agent-platform --keep-stack` — Run with stack kept alive
## Output
Show the user the full command before running, then report test results when done.0f36ad4
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.