我自己回答我的问题。我无法让Jenkins pipeline triggers机制做我想要的事情，无论是pollSCM还是upstream。相反，我通过试验Groovy发现它通过Hudson.instance.queue.getItems()和item.getTask().getName()暴露了所需的信息。后者似乎报告了屏幕上显示的相同构建队列行，例如part of MyJobName #243，从那里我可以看到一个作业是否在构建队列中。
所以，我现在的解决方案是添加一个单独的Jenkins作业，它每10分钟运行一次，从各种repos中检查是否有新的提交，并触发所需的作业。这个触发作业必须在一个特定的单独节点上运行，该节点不运行任何其他长作业，以确保它确实每10分钟运行一次。

pipeline {
    agent {
        label 'node1'
    }
    triggers {
        cron('*/10 * * * *')
    }
    options {
        buildDiscarder(logRotator(numToKeepStr: '10', daysToKeepStr: '1'))
    }
    
    stages {
        stage('Working') {
            steps {
                script {
                    def text = "";
                    for (job in Hudson.instance.queue.getItems()) {
                        text = text.concat("${job.getTask().getName()}\n")
                    }
                    println "Current build queue:\n${text}"
                    writeFile file: 'jenkins-queue.txt', text: "${text}"
                }
                
                sh '''
set +x
URL1='https://github.com/MyCompany/Project1'
URL2='https://github.com/MyCompany/Project2'
URL3='http://svnrepos/code/Projects/Project3/trunk/Production'
URL4='http://svnrepos/code/Projects/Project4/trunk/Production'

BRANCH1='main'
BRANCH2='main'

HASH1=`git ls-remote $URL1 $BRANCH1 | awk '{print $1}'`
HASH2=`git ls-remote $URL2 $BRANCH2 | awk '{print $1}'`
HASH3=`svn info --show-item last-changed-revision $URL3`
HASH4=`svn info --show-item last-changed-revision $URL4`

LINE="$HASH1 $HASH2 $HASH3 $HASH4"
echo "Repo state: $LINE"

if [ -z "$HASH1" ] || [ -z "$HASH2" ] || [ -z "$HASH3" ] || [ -z "$HASH4" ]; then
    echo "Could not retrieve repo state"
    exit 1
fi

JOBS="MyJobName1 MyJobName2 MyJobName3"

rm -f jenkins-trigger.txt
touch jenkins-trigger.txt

for JOB in $JOBS; do
    touch "$JOB.triggered"

    if fgrep -q "$LINE" "$JOB.triggered"; then
        echo "Job already triggered for this repo state: $JOB"
    else
        if egrep -qw "$JOB" jenkins-queue.txt; then
            echo "Job already in Jenkins build queue, marking as triggered: $JOB"
            echo "$LINE" > "$JOB.triggered"
        else
            echo "Triggering job: $JOB"
            echo "$JOB" >> jenkins-trigger.txt
            echo "$LINE" > "$JOB.triggered"
        fi
    fi
done
                '''
                
                script {
                    def text2 = readFile('jenkins-trigger.txt')
                    def lines = text2.readLines()
                    for (line in lines) {                                            
                        println "Triggering job: ${line}"
                        build(job: "${line}", wait: false)
                    }
                }
            }       
        }
    }
}

字符串
当然，中间部分也可以用Groovy编写，我敢肯定。
为了触发级联作业，前一个作业最终可以做类似的事情：

pipeline {
   // ... ...

    post {
        success {
            script {
                def triggerNextJob = true;
                for (job in Hudson.instance.queue.getItems()) {
                    def line = "${job.getTask().getName()}";
                    println "In build queue: ${line}"
                    def pos = line.indexOf("MyNextJobName1");
                    if (pos != -1) {
                        println "Next build already in queue, will not trigger: MyNextJobName1"
                        triggerNextJob = false;
                    }
                }
                if (triggerNextJob) {
                    println "Triggering next job: MyNextJobName1"
                    build(job: "MyNextJobName1", wait: false)
                }
            }
        }
    }
}

型