-
Notifications
You must be signed in to change notification settings - Fork 37
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SWC-6536: handle timing issues related to using larger runner #5184
Conversation
…fail flaky tests more quickly to reduce cost in CI
…form, confirm login was correct before waiting for redirect
// This may be the first test to run and trigger SWC compilation | ||
// ...which may take 15+ seconds. Mark this test as slow, | ||
// ...so the test will be given triple timeout. | ||
// https://playwright.dev/docs/api/class-test#test-slow-1 | ||
setup.slow() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Successful auth.setup
tests are taking 2-4 seconds on the larger Linux runner, 4-6 seconds locally, and 7-11 seconds on the macOS runner. Generally, this test is much faster than the other e2e tests, so stop marking this test as "slow" -- allow the test to fail faster when flaky, so we spend spend less money on larger runner in CI.
const usernameInput = page.getByLabel('Username or Email Address') | ||
await usernameInput.fill(testUserName) | ||
await expect(usernameInput).toHaveValue(testUserName) | ||
|
||
const passwordInput = page.getByLabel('Password') | ||
await expect(passwordInput).toBeEmpty() | ||
await passwordInput.fill(testUserPassword) | ||
await expect(passwordInput).not.toBeEmpty() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ensure that the username/password values have been set before trying to sign in
// Ensure that correct username/password were received | ||
const loadingButton = page.getByRole('button', { name: 'Logging you in' }) | ||
await expect(loadingButton).toBeVisible() | ||
await expect(loadingButton).not.toBeVisible() | ||
await expect( | ||
page.getByText('The provided username/password combination is incorrect'), | ||
).not.toBeVisible() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Confirm that the correct username/password were received, hopefully will help us troubleshoot if login fails to redirect as in this run
@@ -21,7 +21,7 @@ dotenv.config() | |||
export default defineConfig({ | |||
testDir: './e2e', | |||
/* Timeout to allow portal enough time to compile when running locally */ | |||
timeout: 5 * 60 * 1000, | |||
timeout: 2 * 60 * 1000, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reduce overall test timeout, so flaky tests fail faster -- currently, the longest test (teams.loggedin
) takes about 25s locally, 17s on larger runner, and 35-40s on macOS runner, so 2 minutes should be enough time for now. If we add a longer test in the future, we can configure a longer timeout just for that test.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I like the idea of having these fail faster if making these longer doesn't substantially make these less flaky. Optimizing for resilience is good, but I wouldn't worry much about shaving seconds or minutes off to save money.
No description provided.