extend sc tunnel idle timeout to avoid late failures in ui tests

[davidsbailey]

TO DO

• 2025/11/07 16:44:21.958543 [control] [INFO] (server msg) The open file limit in your Operating System is too low. Sauce Labs recommends setting it to at least 16000. Please see https://docs.saucelabs.com/secure-connections/sauce-connect-5/system-requirements/ for more information.

original failure

this PR originally attempted to address an issue where it looks like the sauce connect tunnel shuts down, possibly due to inactivity while ui tests run with chromedriver rather than saucelabs due to the --first-run-local flag. Here is an example failure:

https://drone.cdn-code.org/code-dot-org/code-dot-org/43686/2/5

180 passed. 10 failed. Test count: 190. Duration: 11:21 minutes. Total successful reruns of flaky tests: 0.

https://cucumber-logs.s3.amazonaws.com/circle/43686/Chrome_teacher_tools_authored_hints_output.html?versionId=qnbT33Vi_4tqllxkH1KPF82jCo8fBB3P

NoActiveTunnelForIdentifier - Misconfigured -- No active tunnel found for provided identifier (Selenium::WebDriver::Error::SessionNotCreatedError)

error logging

The original issue could not be debugged, so this PR attempts to add some error logging instead. The current settings are:

code-dot-org/lib/cdo/sauce_connect.rb

Lines 51 to 52 in 193b1d5

	cmd += ["--log-level", "info"]
	cmd += ["--log-http", "proxy:errors"]

In the normal case, you get about 40 lines of output: https://drone.cdn-code.org/code-dot-org/code-dot-org/43796/2/5

In a build with many failures, I saw 20K lines of output: https://drone.cdn-code.org/code-dot-org/code-dot-org/43795/2/5

docs

see: https://docs.saucelabs.com/dev/cli/sauce-connect-5/run/#http-idle-conn-timeout

Testing story

confirmed we are using sc 5 in CI:

[11:44:34] Dave-M4:~/src/cdo (extend-sc-tunnel-timeout)$ bin/drone/shell ec2-XXX-XXX-XXX-XXX.us-west-2.compute.amazonaws.com
ci@d894e3211ad9:/drone/src$ sc version
Version:	 5.2.1

Deployment strategy

Follow-up work