[8.17] (backport #8396) [tests] split up serverless and resource leaks integration tests (#8455)

* [tests] split up serverless and resource leaks integration tests (#8396)

* fix: expose PreinstalledPackages

* fix: make groups.go importable

* feat: introduce new leak package for integration tests

* feat: introduce new serverless package for integration tests

* feat: incorporate serverless and extended runtime leak tests in mage

* fix: normalise powershell (windows) and bash (unix) integration tests invocation flow

* fix: use -test prefixed args for test-wise argument in buildkite-integration-tests.ps1

* fix: add comment about test timeout 2h0m0s

* fix: correct wording

* doc: add integration:TestServerless and update integration:TestForResourceLeaks in test-framework-dev-guide.md

* fix: return directly the test exit code for ps1 scripts

* fix: rename file that host TestMain to main_test.go

* fix: correct build tags for serverless integration tests package

* fix: leverage log.Fatalf in serverless/main_test.go

* feat: expand the BK log group on failure

* tmp: introduce artificial test failure to test buildkite logging group expansion

* Revert "tmp: introduce artificial test failure to test buildkite logging group expansion"

This reverts commit ee6bd63.

* fix: typo

* fix: remove redundant BK log group expansion

(cherry picked from commit 92e139c)

# Conflicts:
#	.buildkite/scripts/buildkite-integration-tests.sh
#	.buildkite/scripts/integration-tests.ps1
#	testing/integration/auditd_monitoring_test.go
#	testing/integration/fleetserver_test.go
#	testing/integration/logs_ingestion_test.go
#	testing/integration/network_traffic_monitoring_test.go
#	testing/integration/osquery_monitoring_test.go

* fix: resolve conflicts

* fix: go fmt

* fix: remove irrelevant files

* fix: remove -integration.fips flag

* fix: point to the correct path for "fleet-server.json" in TestInstallFleetServerBootstrap

---------

Co-authored-by: Panos Koutsovasilis <panos.koutsovasilis@elastic.co>

Author: mergify[bot]

.buildkite/bk.integration.pipeline.yml

@@ -39,10 +39,10 @@ steps:
         depends_on:
           - packaging-windows
         env:
-          TEST_LONG_RUNNING: "true"
+          TEST_PACKAGE: "github.com/elastic/elastic-agent/testing/integration/leak"
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 fleet true TestLongRunningAgentForLeaks
+          .buildkite/scripts/steps/integration_tests_tf.ps1 fleet true
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -57,10 +57,10 @@ steps:
         depends_on:
           - packaging-windows
         env:
-          TEST_LONG_RUNNING: "true"
+          TEST_PACKAGE: "github.com/elastic/elastic-agent/testing/integration/leak"
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 fleet true TestLongRunningAgentForLeaks
+          .buildkite/scripts/steps/integration_tests_tf.ps1 fleet true
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -74,10 +74,10 @@ steps:
       - label: "Ubuntu:2404:amd64:sudo"
         depends_on: packaging-ubuntu-x86-64
         env:
-          TEST_LONG_RUNNING: "true"
+          TEST_PACKAGE: "github.com/elastic/elastic-agent/testing/integration/leak"
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-ubuntu-x86-64'
-          .buildkite/scripts/steps/integration_tests_tf.sh fleet true TestLongRunningAgentForLeaks
+          .buildkite/scripts/steps/integration_tests_tf.sh fleet true
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -99,7 +99,7 @@ steps:
           - packaging-windows
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 {{matrix}} true
+          .buildkite/scripts/steps/integration_tests_tf.ps1 {{matrix}} true
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -124,7 +124,7 @@ steps:
           - packaging-windows
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 {{matrix}} false
+          .buildkite/scripts/steps/integration_tests_tf.ps1 {{matrix}} false
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -143,7 +143,7 @@ steps:
           - packaging-windows
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 {{matrix}} true
+          .buildkite/scripts/steps/integration_tests_tf.ps1 {{matrix}} true
         artifact_paths:
           - build/**
           - build/diagnostics/**
@@ -169,7 +169,7 @@ steps:
           - packaging-windows
         command: |
           buildkite-agent artifact download build/distributions/** . --step 'packaging-windows'
-          .buildkite/scripts/integration-tests.ps1 {{matrix}} false
+          .buildkite/scripts/steps/integration_tests_tf.ps1 {{matrix}} false
         artifact_paths:
           - build/**
           - build/diagnostics/**

.buildkite/integration.pipeline.yml

@@ -96,7 +96,7 @@ steps:
       TEST_INTEG_AUTH_GCP_DATACENTER: "us-central1-a"
     command: |
       buildkite-agent artifact download "build/distributions/**" . $BUILDKITE_BUILD_ID
-      .buildkite/scripts/steps/integration_tests.sh serverless integration:single TestLogIngestionFleetManaged #right now, run a single test in serverless mode as a sort of smoke test, instead of re-running the entire suite
+      .buildkite/scripts/steps/integration_tests.sh serverless integration:testServerless
     artifact_paths:
       - "build/TEST-**"
       - "build/diagnostics/*"

.buildkite/scripts/buildkite-integration-tests.ps1

@@ -0,0 +1,80 @@
+param(
+    [string]$GROUP_NAME,
+    [string]$TEST_SUDO
+)
+
+if (-not $GROUP_NAME)
+{
+    Write-Error "Error: Specify the group name: buildkite-integration-tests.ps1 <group_name> <true|false>"
+    exit 1
+}
+
+if (-not $TEST_SUDO)
+{
+    Write-Error "Error: Specify the test sudo: buildkite-integration-tests.ps1 <group_name> <true|false>"
+    exit 1
+}
+
+if (-not $env:TEST_PACKAGE)
+{
+    $env:TEST_PACKAGE = "github.com/elastic/elastic-agent/testing/integration"
+}
+
+# TODO: make is not available on Windows yet hence we cannot use make install-gotestsum
+go install gotest.tools/gotestsum
+gotestsum --version
+
+$env:TEST_BINARY_NAME = "elastic-agent"
+# Parsing version.go. Will be simplified here: https://github.com/elastic/ingest-dev/issues/4925
+$AGENT_VERSION = (Get-Content version/version.go | Select-String -Pattern 'const defaultBeatVersion =' | ForEach-Object { $_ -replace '.*?"(.*?)".*', '$1' })
+$env:AGENT_VERSION = $AGENT_VERSION + "-SNAPSHOT"
+
+Write-Output "~~~ Agent version: $env:AGENT_VERSION"
+$env:SNAPSHOT = $true
+
+Write-Host "~~~ Running integration tests as $env:USERNAME"
+
+$osInfo = (Get-CimInstance Win32_OperatingSystem).Caption + " " + (Get-CimInstance Win32_OperatingSystem).OSArchitecture -replace " ", "_"
+$root_suffix = ""
+if ($TEST_SUDO -eq "true")
+{
+    $root_suffix = "_sudo"
+}
+$fully_qualified_group_name = "${GROUP_NAME}${root_suffix}_${osInfo}"
+$outputXML = "build/${fully_qualified_group_name}.integration.xml"
+$outputJSON = "build/${fully_qualified_group_name}.integration.out.json"
+
+$TestsExitCode = 0
+
+try
+{
+    Write-Output "~~~ Integration tests: $GROUP_NAME as user: $env:USERNAME"
+    # -test.timeout=2h0m0s is set because some tests normally take up to 45 minutes.
+    # This 2-hour timeout provides enough room for future, potentially longer tests,
+    # while still enforcing a reasonable upper limit on total execution time.
+    # See: https://pkg.go.dev/cmd/go#hdr-Testing_flags
+    $gotestArgs = @("-tags=integration", "-test.shuffle=on", "-test.timeout=2h0m0s", "$env:TEST_PACKAGE", "-v", "-args", "-integration.groups=$GROUP_NAME", "-integration.sudo=$TEST_SUDO")
+    & gotestsum --no-color -f standard-quiet --junitfile-hide-skipped-tests --junitfile "${outputXML}" --jsonfile "${outputJSON}" -- @gotestArgs
+    $TestsExitCode = $LASTEXITCODE
+
+    if ($TestsExitCode -ne 0)
+    {
+        Write-Output "^^^ +++"
+        Write-Output "Integration tests failed"
+    }
+}
+finally
+{
+    if (Test-Path $outputXML)
+    {
+        # Install junit2html if not installed
+        go install github.com/alexec/junit2html@latest
+        Get-Content $outputXML | junit2html > "build/TEST-report.html"
+    }
+    else
+    {
+        Write-Output "Cannot generate HTML test report: $outputXML not found"
+    }
+}
+
+exit $TestsExitCode

.buildkite/scripts/buildkite-integration-tests.sh

@@ -3,7 +3,9 @@
 
 GROUP_NAME=$1
 TEST_SUDO=$2
-TEST_NAME_PATTERN=${3:-""}
+
+# Set default TEST_PACKAGE if not already defined in env or argument
+: "${TEST_PACKAGE:="github.com/elastic/elastic-agent/testing/integration"}"
 
 if [ -z "$GROUP_NAME" ]; then
   echo "Error: Specify the group name: sudo-integration-tests.sh [group_name]" >&2
@@ -46,18 +48,22 @@ outputXML="build/${fully_qualified_group_name}.integration.xml"
 outputJSON="build/${fully_qualified_group_name}.integration.out.json"
 
 echo "~~~ Integration tests: ${GROUP_NAME}"
-GOTEST_ARGS=(-tags integration -test.shuffle on -test.timeout 2h0m0s)
-if [ -n "$TEST_NAME_PATTERN" ]; then
-  GOTEST_ARGS+=(-run="${TEST_NAME_PATTERN}")
-fi
-GOTEST_ARGS+=("github.com/elastic/elastic-agent/testing/integration" -v -args "-integration.groups=${GROUP_NAME}" "-integration.sudo=${TEST_SUDO}")
-
+# -test.timeout=2h0m0s is set because some tests normally take up to 45 minutes.
+# This 2-hour timeout provides enough room for future, potentially longer tests,
+# while still enforcing a reasonable upper limit on total execution time.
+# See: https://pkg.go.dev/cmd/go#hdr-Testing_flags
+GOTEST_ARGS=(-tags integration -test.shuffle on -test.timeout 2h0m0s "${TEST_PACKAGE}" -v -args "-integration.groups=${GROUP_NAME}" "-integration.sudo=${TEST_SUDO}")
 set +e
 TEST_BINARY_NAME="elastic-agent" AGENT_VERSION="${AGENT_VERSION}" SNAPSHOT=true \
   gotestsum --no-color -f standard-quiet --junitfile "${outputXML}" --junitfile-hide-skipped-tests --jsonfile "${outputJSON}" -- "${GOTEST_ARGS[@]}"
 TESTS_EXIT_STATUS=$?
 set -e
 
+if [[ $TESTS_EXIT_STATUS -ne 0 ]]; then
+   echo "^^^ +++"
+   echo "Integration tests failed"
+fi
+
 if [ -f "$outputXML" ]; then
   go install github.com/alexec/junit2html@latest
   junit2html < "$outputXML" > build/TEST-report.html

.buildkite/scripts/steps/integration_tests_tf.ps1

@@ -0,0 +1,43 @@
+param (
+    [string]$GROUP_NAME,
+    [string]$TEST_SUDO
+)
+
+Write-Output "~~~ Preparing environment"
+
+$PSVersionTable.PSVersion
+
+. "$PWD\.buildkite\scripts\steps\ess.ps1"
+
+# Read package version from .package-version file
+$PACKAGE_VERSION = Get-Content .package-version -ErrorAction SilentlyContinue
+if ($PACKAGE_VERSION) {
+    $PACKAGE_VERSION = "${PACKAGE_VERSION}-SNAPSHOT"
+}
+
+Write-Output "~~~ Building test binaries"
+& mage build:testBinaries
+if ($LASTEXITCODE -ne 0) {
+    Write-Output "^^^ +++"
+    Write-Output "Failed to build test binaries"
+    exit 1
+}
+
+$TestsExitCode = 0
+try {
+    Write-Output "~~~ Running integration tests"
+    # Get-Ess-Stack will start the ESS stack if it is a BK retry, otherwise it will retrieve ESS stack metadata
+    Get-Ess-Stack -StackVersion $PACKAGE_VERSION
+    & "$PWD\.buildkite\scripts\buildkite-integration-tests.ps1" $GROUP_NAME $TEST_SUDO
+    $TestsExitCode = $LASTEXITCODE
+    if ($TestsExitCode -ne 0)
+    {
+        Write-Output "^^^ +++"
+        Write-Output "Integration tests failed"
+    }
+} finally {
+    # ess_down will destroy the ESS stack if tf state file is found, aka if this is a BK retry
+    ess_down
+}
+
+exit $TestsExitCode

.buildkite/scripts/steps/integration_tests_tf.sh

@@ -9,9 +9,7 @@ asdf install
 
 GROUP_NAME=$1
 TEST_SUDO=$2
-# NOTE: This argument is not used in this script, but is declared to show that it can be set
-# and passed down to downstream scripts where it may be used.
-TEST_NAME_PATTERN=${3:-""}
+
 if [ -z "$GROUP_NAME" ]; then
   echo "Error: Specify the group name: integration_tests_tf.sh [group_name]" >&2
   exit 1
@@ -37,7 +35,7 @@ mage build:testBinaries
 if [[ "${BUILDKITE_RETRY_COUNT}" -gt 0 ]]; then
   echo "~~~ The steps is retried, starting the ESS stack again"
   trap 'ess_down' EXIT
-  ess_up $OVERRIDE_STACK_VERSION || echo "Failed to start ESS stack" >&2
+  ess_up $OVERRIDE_STACK_VERSION || (echo -e "^^^ +++\nFailed to start ESS stack")
   preinstall_fleet_packages
 else
   # For the first run, we start the stack in the start_ess.sh step and it sets the meta-data

docs/test-framework-dev-guide.md

@@ -130,10 +130,13 @@ We pass a `-test.count` flag along with the name match
 We pass a `-test.run` flag along with the names of the tests we want to run in OR
 `GOTEST_FLAGS="-test.run ^(TestStandaloneUpgrade|TestFleetManagedUpgrade)$" mage integration:test`
 
+##### Run Serverless tests
+The test framework includes a smoke test suite to check elastic-agent in a serverless environment. The suite can be run via the `integration:TestServerless` mage target.
+
 ##### Run Extended Runtime Leak Test
 The test framework includes a "long running" test to check for resource leaks and stability.
 The runtime of the test can be set via the `LONG_TEST_RUNTIME` environment variable.
-The test itself can be run via the `integration:TestLongRunningAgentForLeaks` mage target.
+The test itself can be run via the `integration:TestForResourceLeaks` mage target.
 
 ##### Limitations
 Due to the way the parameters are passed to `devtools.GoTest` the value of the environment variable