[9.3] (backport #14836) Use the checkin on state change mode in integration tests

[mergify]

• Relates Add support for quicker Fleet check-in upon component status #9982

Uses the checkin on state change mode in integration tests, to speed up tests that poll for the agent status to change in Fleet. This makes a few test significantly faster, like the log level related tests.

elastic-agent/testing/integration/ess/log_level_test.go

Lines 263 to 272 in 9430866

	// assert Fleet eventually receives the new log level from agent through checkin
	assert.Eventuallyf(t, func() bool {
	fleetMetadataLogLevel, err := getLogLevelFromFleetMetadata(ctx, info.KibanaClient, agentID)
	if err != nil {
	t.Logf("error getting log level for agent %q from Fleet metadata: %v", agentID, err)
	return false
	}
	t.Logf("Fleet metadata log level for agent %q: %q policy log level: %q", agentID, fleetMetadataLogLevel, policyLogLevel)
	return fleetMetadataLogLevel == policyLogLevel.String()
	}, 6*time.Minute, 30*time.Second, "agent never communicated policy log level %q to Fleet", policyLogLevel)

e
The impact on test execution overall is not that big because these tests weren't on the critical path. It's also bypassed in several of the upgrade tests until I backport this and come back and change the version guard.

---

This is an automatic backport of pull request #14836 done by Mergify.

[mergify]

Cherry-pick of cb4ff55 has failed:

On branch mergify/bp/9.3/pr-14836
Your branch is up to date with 'origin/9.3'.

You are currently cherry-picking commit cb4ff5574.
  (fix conflicts and run "git cherry-pick --continue")
  (use "git cherry-pick --skip" to skip this patch)
  (use "git cherry-pick --abort" to cancel the cherry-pick operation)

Changes to be committed:
	modified:   internal/pkg/agent/application/enroll/enroll.go
	modified:   internal/pkg/agent/application/enroll/options.go
	modified:   internal/pkg/agent/cmd/enroll.go
	modified:   internal/pkg/agent/cmd/enroll_cmd_test.go
	modified:   pkg/testing/fixture.go
	modified:   pkg/testing/fixture_install.go
	modified:   testing/integration/ess/log_level_test.go

Unmerged paths:
  (use "git add <file>..." to mark resolution)
	both modified:   internal/pkg/agent/configuration/fleet.go

To fix up this pull request, you can check it out locally. See documentation: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/reviewing-changes-in-pull-requests/checking-out-pull-requests-locally

[github-actions]

TL;DR

golangci-lint failed in lint (ubuntu-latest) on 5 new lint violations (gosec, noctx, nolintlint, and two unused constants). Fix those specific findings and re-run the workflow.

Remediation

• Update pkg/testing/fixture_install.go to avoid opening callback-derived walk paths in a way that triggers gosec G122 (use a root-scoped/symlink-safe access pattern in that walk path handling).
• In internal/pkg/agent/cmd/enroll_cmd_test.go, replace net.Listen(...) with (&net.ListenConfig{}).Listen(context.Background(), ...), and remove the now-unused //nolint:gosec directive on the TLS config line.
• In internal/pkg/agent/configuration/fleet.go, either remove fleetCheckinModeStandard and fleetCheckinModeOnStateChanged or wire them into live code so they are no longer unused.

Investigation details

Root Cause

The run failed in the golangci-lint step due to five lint violations introduced in the PR patch. This is materially different from the prior detective report on this PR (which flagged missing checkin-compression symbols).

Evidence

• Workflow: https://github.com/elastic/elastic-agent/actions/runs/27185962393
• Job/step: lint (ubuntu-latest) → golangci-lint (https://github.com/elastic/elastic-agent/actions/runs/27185962393/job/80254844060)
• Key log excerpt:

pkg/testing/fixture_install.go:872:23: G122 ... race-prone path ... (gosec)
internal/pkg/agent/cmd/enroll_cmd_test.go:925:30: net.Listen must not be called ... (noctx)
internal/pkg/agent/cmd/enroll_cmd_test.go:995:22: directive `//nolint:gosec ...` is unused ... (nolintlint)
internal/pkg/agent/configuration/fleet.go:100:2: const fleetCheckinModeStandard is unused (unused)
internal/pkg/agent/configuration/fleet.go:101:2: const fleetCheckinModeOnStateChanged is unused (unused)
issues found

Validation

• Not run locally (detective workflow is read-only); diagnosis is from GitHub Actions job logs.

Follow-up

• After applying the above lint fixes, re-run golangci-lint/PR checks to confirm no additional --new-from-patch issues remain.

---

What is this? | From workflow: PR Actions Detective

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.

[elasticmachine]

💔 Build Failed

• Buildkite Build
• Commit: 6bbb3b4

Failed CI Steps

• Unit tests - Ubuntu 22.04
• Unit tests - Ubuntu 22.04
• Unit tests - Ubuntu 22.04 with requirefips build tag
• Unit tests - Ubuntu 22.04 with requirefips build tag
• Unit tests - Ubuntu 22.04 ARM64
• Unit tests - Ubuntu 22.04 ARM64
• Unit tests - Windows 2022
• Unit tests - Windows 2022
• Unit tests - Windows 2016
• Unit tests - Windows 2016
• Unit tests - macOS 15 ARM
• Unit tests - macOS 15 ARM
• Unit tests - Windows 10
• Unit tests - Windows 10
• Unit tests - Windows 11
• Unit tests - Windows 11
• Unit tests - Windows 11 arm64
• Unit tests - Windows 11 arm64
• windows:amd64:tier1:sudo:fleet
• linux:amd64:tier1:sudo:fqdn
• linux:amd64:tier2:sudo:fqdn:platform-ingest-elastic-agent-debian-13-1780275718

History

• 💔 Build #41230 failed cd67856
• 💔 Build #41224 failed 6446190

cc @cmacknz

[github-actions]

TL;DR

Buildkite unit tests are failing due to a nil-pointer panic introduced in the new checkin-on-state-change enroll path. Initialize cfg.Checkin before writing cfg.Checkin.Mode in createFleetConfigFromEnroll.

Remediation

• In internal/pkg/agent/application/enroll/enroll.go, guard/initialize cfg.Checkin before cfg.Checkin.Mode = configuration.FleetCheckinModeOnStateChange (for example: if cfg.Checkin == nil { cfg.Checkin = &configuration.FleetCheckin{} }).
• Re-run the failing Unit tests steps (especially Unit tests - Ubuntu 22.04 with requirefips build tag) to confirm the enroll tests no longer panic.

Investigation details

Root Cause

This is a code bug. The PR adds checkin-mode wiring in enroll:

• internal/pkg/agent/application/enroll/enroll.go:500 writes cfg.Checkin.Mode = ... when checkinOnStateChange is true.

But on this backport branch, the default fleet config sets checkin to nil:

• internal/pkg/agent/configuration/fleet.go:62 has Checkin: nil in DefaultFleetAgentConfig().

So cfg.Checkin.Mode dereferences a nil pointer and panics.

Evidence

• Build: https://buildkite.com/elastic/elastic-agent/builds/41243
• Job/step: Unit tests - Ubuntu 22.04 with requirefips build tag (.buildkite/scripts/steps/unit-tests.sh)
• Key log excerpt:

=== FAIL: internal/pkg/agent/cmd TestEnroll/checkin-on-state-change_flag_configures_on_state_change_checkin_mode
panic: runtime error: invalid memory address or nil pointer dereference
...
internal/pkg/agent/application/enroll/enroll.go:500 +0x5c9

• Diff context also shows the newly introduced dereference:
  • internal/pkg/agent/application/enroll/enroll.go hunk @@ -490,12 +490,15 @@ adds:

if checkinOnStateChange {
    cfg.Checkin.Mode = configuration.FleetCheckinModeOnStateChange
}

Verification

• Not run locally (detective workflow is read-only); diagnosis is based on Buildkite logs plus PR branch source at commit 6bbb3b4d5fc9952ae4cef1029f224ceed925d4c0.

Follow-up

• Checked for a matching flaky-test issue for TestEnroll/checkin-on-state-change_flag_configures_on_state_change_checkin_mode; none found.

---

What is this? | From workflow: PR Buildkite Detective

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.