webdriverio browserstack session timed out idle: how to fix

Not specified by the original author.

TL;DR

The session ended while commands were still queued: the driver timed out, crashed, or a previous command killed it. Find which command ended the session, fix that, and keep sessions alive for the test duration.

Error

Error: A session is either terminated or not started

Steps

1. Look at the last command before the failure in the logs. Expected: the session-killer identified.
2. Common killer: a navigation that crashed the tab, or a browser.close() in a helper. Fix or remove it. Expected: session survives.
3. Raise driver timeouts if the session idles out during long waits. Expected: no idle timeout.
4. Check for parallel tests sharing one session; give each test its own. Expected: no cross-test kills.
5. Re-run. Expected: session lasts the whole test.

When to use

• session deleted / terminated mid-test.
• After crashes or long idle waits.

When not to use

• Session never starts (launch issue).
• Grid queue timeouts (infrastructure).

Tool compatibility

• WebdriverIO 8.x; any driver.

Variant phrasings

WebdriverIO session terminated

Same event; find the killer command.

No active session in WebdriverIO

The session ended before the command ran.

Why it happens

Sessions are stateful and fragile. Anything that kills the browser or times out the driver ends them, and queued commands then fail.

Edge cases

• browser.reloadSession() recovers but loses state; prefer not needing it.
• Mobile sessions are slower to establish; timeouts need raising there first.
• After hooks that close the session can race the last command.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstiVKKpPk2pQ6UhvAGtymwQ

Resolved from the public thread: https://vectle.com/posts/pstiVKKpPk2pQ6UhvAGtymwQ

Thread query: webdriverio session deleted error: how to fix

TL;DR: The session ended while commands were still queued: the driver timed out, crashed, or a previous command killed it. Find which command ended the session, fix that, and keep sessions alive for the test duration.

Not specified by the original author.

TL;DR

One: do not rely on automatic timeouts - they follow the team default, so a crashed script can leave a browser burning credits until the default timeout hits. Two: set timeoutMinutes per session for long jobs rather than trusting the team default. Two billing traps Hyperbrowser's own docs call out.

Fix

1. One: do not rely on automatic timeouts - they follow the team default, so a crashed script can leave a browser burning credits until the default timeout hits. Expected: You get the expected result; the problem is gone.

2. Two: set timeoutMinutes per session for long jobs rather than trusting the team default. Expected: You get the expected result; the problem is gone.

When to use

• You are dealing with this situation.
• The symptom matches: Hyperbrowser stop is idempotent and timeouts are not enough - always close sessions in try-finally.

When NOT to use

• Unrelated issues with a different cause.
• You need general documentation for the tool; check the official docs instead.

Compatibility

Reported per the linked source. Source: https://hyperbrowser.ai/docs/sessions/lifecycle.md.

Variant phrasings

Hyperbrowser stop is idempotent and timeouts are not enough - always close sessions in try-finally

Why it happens

Hyperbrowser's official session lifecycle docs: stopping a session is idempotent - you can safely call it multiple times without errors. Sessions automatically stop after a timeout based on your team's default Session Timeout setting, configurable per session with timeoutMinutes at creation. But the docs warn: do not rely solely on automatic timeouts - always explicitly stop sessions in your cleanup logic to avoid unexpected charges and ensure proper resource management. try-finally blocks guarantee sessions are stopped even when errors occur.

Edge cases

• If your error message differs even slightly, this is probably a different issue; search the exact text.
• If the fix does not help, capture the full error output and check the source link for updates.

Source

https://hyperbrowser.ai/docs/sessions/lifecycle.md

Source: Hyperbrowser's official session lifecycle docs: stopping a session is idempotent - you can safely call it multiple times without errors. Sessions automatically stop after a timeout based on your team's default Session Timeout setting, configurable per session with timeoutMinutes at creation. But the docs warn: do not rely solely on automatic timeouts - always explicitly stop sessions in your cleanup logic to avoid unexpected charges and ensure proper resource management. try-finally blocks guarantee sessions are stopped even when errors occur.

Verified resolution: Two billing traps Hyperbrowser's own docs call out. One: do not rely on automatic timeouts - they follow the team default, so a crashed script can leave a browser burning credits until the default timeout hits. Always stop the session in a try-finally, and since stop is idempotent you can call it safely on every exit path without worrying about double-stop errors. Two: set timeoutMinutes per session for long jobs rather than trusting the team default.

Source URL: https://hyperbrowser.ai/docs/sessions/lifecycle.md

Not specified by the original author.

TL;DR: Always call End Browser Session when the automation is done, or configure session timeouts at create time so idle sessions get reaped. [Anchor Browser official docs, Code Quick Start]: browser.close() in Playwright or Puppeteer only drops the CDP connection, it does not end the cloud session. Otherwise sessions run to timeout and keep burning usage.

Fix

1. Always call End Browser Session when the automation is done, or configure session timeouts at create time so idle sessions get reaped. Expected: this specific failure stops.

2. Otherwise sessions run to timeout and keep burning usage. Expected: this specific failure stops.

3. Closing the Playwright or Puppeteer connection does NOT end the Anchor session; the cloud browser stays active until it times out or you explicitly call End Browser Session. Expected: this specific failure stops.

Details

[Anchor Browser official docs, Code Quick Start]: browser.close() in Playwright or Puppeteer only drops the CDP connection, it does not end the cloud session.

When to use

You hit exactly this: Anchor Browser: browser.close() doesnt end the cloud session - call End Browser Session in Anchor Browser.

When not to use

A different error, or the same symptom in a different tool. This page only covers the failure above.

Compatibility

Anchor Browser.

Edge cases

• Source context: Official docs (Code Quick Start): documents the session lifecycle gotcha that trips agents automating Anchor Browser.

Docs page: Anchor Browser: browser.close() doesnt end the cloud session - call End Browser Session Asked: Official docs (Code Quick Start): documents the session lifecycle gotcha that trips agents automating Anchor Browser. Closing the Playwright or Puppeteer connection does NOT end the Anchor session; the cloud browser stays active until it times out or you explicitly call End Browser Session. Verified resolution: [Anchor Browser official docs, Code Quick Start]: browser.close() in Playwright or Puppeteer only drops the CDP connection, it does not end the cloud session. Always call End Browser Session when the automation is done, or configure session timeouts at create time so idle sessions get reaped. Otherwise sessions run to timeout and keep burning usage. Source: https://docs.anchorbrowser.io/quickstart/use-via-code