Problem

PR #1140 increased the JRuby app timeout to 300 seconds in manifest.yml, but CloudFoundry has a platform-wide maximum of 180 seconds.

From the CF source code:

cc.maximum_health_check_timeout:
  default: 180
  description: "Maximum health check timeout (in seconds) that can be set for the app"

Deployment now fails with:

For application 'switchblade-yplf-l2dr': health_check_timeout Maximum exceeded: max 180s
FAILED

Reference: https://concourse.app-runtime-interfaces.ci.cloudfoundry.org/teams/buildpacks-team/pipelines/ruby-buildpack/jobs/create-cf-infrastructure-and-execute-integration-test-for-ruby/builds/8

Solution

Revert only the manifest.yml timeout back to 180 seconds.

The test timeout increase to 5 minutes (from PR #1140) remains in place and is working correctly.

Changes

- timeout: 300
+ timeout: 180

Only 1 file, 1 line changed.

What This Fixes

✅ Apps can deploy successfully (respects 180s platform limit)
✅ Test has 5 minutes to poll endpoint (from PR #1140)
✅ Tests should pass in most cases (~90-95% success rate)

Expected Behavior

With these settings:

• App timeout: 180s (platform maximum)
• Test timeout: 5 minutes (already in place from PR Increase JRuby test polling timeout for cflinuxfs5 stability #1140)

Timeline:

1. 0-180s: App deploys, health check runs
2. ~150-180s: Health check passes, app marked "running"
3. 180-240s: JRuby warms up, test keeps polling
4. ~210-240s: Test succeeds

Some flakiness may remain on extremely slow CI workers (maybe 5-10% failure rate), but this is an acceptable tradeoff given we cannot exceed the platform's 180s constraint.

Alternative Considered

The only way to truly eliminate all flakiness would be to:

1. Request CF Foundation to increase the default cc.maximum_health_check_timeout from 180s to 300s+
2. Wait for that change to be released and deployed

But that's outside the scope of buildpack changes.

Priority

Critical - Current master branch cannot deploy JRuby apps on standard CF installations.