Check Jenkins job status after triggering a build remotely
I had similar problem to get state with rest api only.
this was my solution (it is a weak and not stable solution!):
#Ex. http://jenkins.com/job/test
JOB_URL="${JENKINS_URL}/job/${JOB_NAME}"
#here you can ask for next build job number
function getNextBuildNr(){
curl --silent ${JOB_URL}/api/json | grep -Po '"nextBuildNumber":\K\d+'
}
# this will request the Status of job
function getBuildState(){
buildNr=$1
curl --silent ${JOB_URL}/${buildNr}/api/json | grep -Po '"result":\s*"\K\w+'
}
#this will wait for your Job state, by polling Jenkins every second
function waitForJob() {
buildNr=$1
state=""
while [ "${state}" == "" ]
do
sleep 1
state=$(getBuildState ${buildNr})
echo -n '.'
done
echo -e "\n"
}
#now you can run and build
BUILD_NR=$(getNextBuildNr)
# input here your code/function to trigger the job
waitForJob ${BUILD_NR}
BUILD_STATE=$(getBuildState ${BUILD_NR})
echo "$BUILD_STATE"
You can use Jenkins API for this. A sample Python script:
import json
import requests
import time
job_name = "testjob" . #Give your job name here
def jenkins_job_status(job_name):
try:
url = "https://your_jenkins_endpoint/job/%s/lastBuild/api/json" %job_name #Replace 'your_jenkins_endpoint' with your Jenkins URL
while True:
data = requests.get(url).json()
if data['building']:
time.sleep(60)
else:
if data['result'] == "SUCCESS":
print "Job is success"
return True
else:
print "Job status failed"
return False
except Exception as e:
print str(e)
return False
if __name__ == "__main__":
if jenkins_job_status(job_name):
print "Put your autmation here for 'job is success' condition"
else:
print "Put your autmation here for 'job is failed' condition"
Refer http://www.easyaslinux.com/tutorials/devops/how-to-check-build-status-of-jenkins-job-using-python-script/ for detailed explanation
I solved this problem using polling of the Jenkins server. When a job is started remotely the return headers has the job queue URL. Using this one can make more API calls to get status.
Steps:
- start the job
- parse return 'Location' header
- poll the queue looking for job to start
- job queue entry will have an 'executable' entry in its json or xml with job number once it starts
- poll the job status waiting for a result
I used python and the Requests module to do this
#!/usr/bin/python
import requests
import re
import sys
import json
import time
# secs for polling Jenkins API
#
QUEUE_POLL_INTERVAL = 2
JOB_POLL_INTERVAL = 20
OVERALL_TIMEOUT = 3600 # 1 hour
# job specifics: should be passed in
auth_token = 'buildmaster:173223588624f980c3AAA68d4d8efe0c'
jenkins_uri = '192.168.115.187:8080'
job_name = 'rf_systest'
build_token = 'rf_systest_auth_token'
# start the build
#
start_build_url = 'http://{}@{}/job/{}/build?token={}'.format(
auth_token, jenkins_uri, job_name, build_token)
r = requests.post(start_build_url)
# from return headers get job queue location
#
m = re.match(r"http.+(queue.+)\/", r.headers['Location'])
if not m:
# To Do: handle error
print "Job starte request did not have queue location"
sys.exit(1)
# poll the queue looking for job to start
#
queue_id = m.group(1)
job_info_url = 'http://{}@{}/{}/api/json'.format(auth_token, jenkins_uri, queue_id)
elasped_time = 0
print '{} Job {} added to queue: {}'.format(time.ctime(), job_name, job_info_url)
while True:
l = requests.get(job_info_url)
jqe = l.json()
task = jqe['task']['name']
try:
job_id = jqe['executable']['number']
break
except:
#print "no job ID yet for build: {}".format(task)
time.sleep(QUEUE_POLL_INTERVAL)
elasped_time += QUEUE_POLL_INTERVAL
if (elasped_time % (QUEUE_POLL_INTERVAL * 10)) == 0:
print "{}: Job {} not started yet from {}".format(time.ctime(), job_name, queue_id)
# poll job status waiting for a result
#
job_url = 'http://{}@{}/job/{}/{}/api/json'.format(auth_token, jenkins_uri, job_name, job_id)
start_epoch = int(time.time())
while True:
print "{}: Job started URL: {}".format(time.ctime(), job_url)
j = requests.get(job_url)
jje = j.json()
result = jje['result']
if result == 'SUCCESS':
# Do success steps
print "{}: Job: {} Status: {}".format(time.ctime(), job_name, result)
break
elif result == 'FAILURE':
# Do failure steps
print "{}: Job: {} Status: {}".format(time.ctime(), job_name, result)
break
elif result == 'ABORTED':
# Do aborted steps
print "{}: Job: {} Status: {}".format(time.ctime(), job_name, result)
break
else:
print "{}: Job: {} Status: {}. Polling again in {} secs".format(
time.ctime(), job_name, result, JOB_POLL_INTERVAL)
cur_epoch = int(time.time())
if (cur_epoch - start_epoch) > OVERALL_TIMEOUT:
print "{}: No status before timeout of {} secs".format(OVERALL_TIMEOUT)
sys.exit(1)
time.sleep(JOB_POLL_INTERVAL)
Output:
Tue Jan 30 16:24:08 2018: Job rf_systest added to queue: http://buildmaster:[email protected]:8080/queue/item/164/api/json
Tue Jan 30 16:24:39 2018: Job rf_systest not started yet from queue/item/164
Tue Jan 30 16:25:00 2018: Job started URL: http://buildmaster:[email protected]:8080/job/rf_systest/79/api/json
Tue Jan 30 16:25:01 2018: Job: rf_systest Status: None. Polling again in 20 secs
Tue Jan 30 16:25:21 2018: Job started URL: http://buildmaster:[email protected]:8080/job/rf_systest/79/api/json
Tue Jan 30 16:25:21 2018: Job: rf_systest Status: None. Polling again in 20 secs
Tue Jan 30 16:25:41 2018: Job started URL: http://buildmaster:[email protected]:8080/job/rf_systest/79/api/json
Tue Jan 30 16:25:41 2018: Job: rf_systest Status: None. Polling again in 20 secs
Tue Jan 30 16:26:01 2018: Job started URL: http://buildmaster:[email protected]:8080/job/rf_systest/79/api/json
Tue Jan 30 16:26:01 2018: Job: rf_systest Status: None. Polling again in 20 secs
Tue Jan 30 16:26:21 2018: Job started URL: http://buildmaster:[email protected]:8080/job/rf_systest/79/api/json
Tue Jan 30 16:26:21 2018: Job: rf_systest Status: SUCCESS
JSON from a Jenkins queue once its job has started:
{
"_class": "hudson.model.Queue$LeftItem",
"actions": [
{
"_class": "hudson.model.CauseAction",
"causes": [
{
"_class": "hudson.model.Cause$RemoteCause",
"addr": "10.20.30.60",
"note": null,
"shortDescription": "Started by remote host 10.20.30.60"
}
]
}
],
"blocked": false,
"buildable": false,
"cancelled": false,
"executable": {
"_class": "org.jenkinsci.plugins.workflow.job.WorkflowRun",
"number": 45,
"url": "http://192.168.115.187:8080/job/rf_systest/45/"
},
"id": 95,
"inQueueSince": 1517342648136,
"params": "",
"stuck": false,
"task": {
"_class": "org.jenkinsci.plugins.workflow.job.WorkflowJob",
"color": "blue_anime",
"name": "rf_systest",
"url": "http://192.168.115.187:8080/job/rf_systest/"
},
"url": "queue/item/95/",
"why": null
}
When you trigger a job, the job is placed into the queue. The actual build is created only when it starts running and at that point the build gets a build number. If all your executors are busy, it can sometimes take a long time before the build is created and starts running.
The only way to get the build number when triggering a job, is to use the "build" command of the Jenkins CLI. If you use the -w option, the command will not return until the build starts and then it will print "Started build #N"
You do not actually need the java cli.jar, just an ssh client is enough. See https://wiki.jenkins-ci.org/display/JENKINS/Jenkins+SSH
Other than that there is no known solution. You might be able to search through the builds and find a one that was triggered around the time your triggered the job, but that's a lot of work.