Use schema last cache to speed up the last and last_by aggregations in table model #23887

	name: Cluster IT - 1C3D

	on:
	push:
	branches:
	- master
	- 'rel/1.*'
	- 'rc/1.*'
	paths-ignore:
	- 'docs/**'
	- 'site/**'
	pull_request:
	branches:
	- master
	- 'rel/1.*'
	- 'rc/1.*'
	paths-ignore:
	- 'docs/**'
	- 'site/**'
	# allow manually run the action:
	workflow_dispatch:

	concurrency:
	group: ${{ github.workflow }}-${{ github.ref }}
	cancel-in-progress: true

	env:
	MAVEN_OPTS: -Dhttp.keepAlive=false -Dmaven.wagon.http.pool=false -Dmaven.wagon.http.retryHandler.class=standard -Dmaven.wagon.http.retryHandler.count=3
	MAVEN_ARGS: --batch-mode --no-transfer-progress
	DEVELOCITY_ACCESS_KEY: ${{ secrets.GE_ACCESS_TOKEN }}

	jobs:
	Simple:
	strategy:
	fail-fast: false
	max-parallel: 20
	matrix:
	java: [ 17 ]
	runs-on: [self-hosted, iotdb]
	# group: self-hosted
	# labels: iotdb
	steps:
	- uses: actions/checkout@v4
	- name: Set up JDK ${{ matrix.java }}
	uses: actions/setup-java@v4
	with:
	distribution: liberica
	java-version: ${{ matrix.java }}
	- name: IT/UT Test
	shell: bash
	# we do not compile client-cpp for saving time, it is tested in client.yml
	# we can skip influxdb-protocol because it has been tested separately in influxdb-protocol.yml
	run: \|
	mvn clean verify \
	-P with-integration-tests \
	-DskipUTs \
	-DintegrationTest.forkCount=6 -DConfigNodeMaxHeapSize=1024 -DDataNodeMaxHeapSize=1024 \
	-pl integration-test \
	-am -PClusterIT
	- name: Upload Artifact
	if: failure()
	uses: actions/upload-artifact@v4
	with:
	name: cluster-log-java${{ matrix.java }}-${{ runner.os }}
	path: integration-test/target/cluster-logs
	retention-days: 1

Provide feedback