pipeline {
/* insert Declarative Pipeline here */
}
This section builds on the information introduced in Getting started with Pipeline and should be treated solely as a reference. For more information on how to use Pipeline syntax in practical examples, refer to the Using a Jenkinsfile section of this chapter. As of version 2.5 of the Pipeline plugin, Pipeline supports two discrete syntaxes which are detailed below. For the pros and cons of each, refer to the Syntax Comparison.
As discussed at the start of this chapter, the most fundamental part of a Pipeline is the "step". Basically, steps tell Jenkins what to do and serve as the basic building block for both Declarative and Scripted Pipeline syntax.
For an overview of available steps, please refer to the Pipeline Steps reference which contains a comprehensive list of steps built into Pipeline as well as steps provided by plugins.
Declarative Pipeline presents a more simplified and opinionated syntax on top of the Pipeline sub-systems.
All valid Declarative Pipelines must be enclosed within a pipeline
block, for example:
pipeline {
/* insert Declarative Pipeline here */
}
The basic statements and expressions which are valid in Declarative Pipeline follow the same rules as Groovy’s syntax with the following exceptions:
The top-level of the Pipeline must be a block, specifically: pipeline { }
.
No semicolons as statement separators. Each statement has to be on its own line.
Blocks must only consist of Sections, Directives, Steps, or assignment statements.
A property reference statement is treated as a no-argument method invocation.
So, for example, input
is treated as input()
.
You can use the Declarative Directive Generator to help you get started with configuring the directives and sections in your Declarative Pipeline.
There is currently an open issue which limits the maximum size of the code within the pipeline{}
block.
This limitation does not apply to Scripted Pipelines.
Sections in Declarative Pipeline typically contain one or more Directives or Steps.
The agent
section specifies where the entire Pipeline, or a specific stage, will execute in the Jenkins environment depending on where the agent
section is placed.
The section must be defined at the top-level inside the pipeline
block, but stage-level usage is optional.
Required |
Yes |
---|---|
Parameters |
|
Allowed |
In the top-level |
There are some nuances when adding an agent to the top level or a stage level when the options
directive is applied.
Check the section options for more information.
In agents
declared at the top level of a Pipeline, an agent is allocated and then the timeout
option is applied.
The time to allocate the agent is not included in the limit set by the timeout
option.
pipeline {
agent any
options {
// Timeout counter starts AFTER agent is allocated
timeout(time: 1, unit: 'SECONDS')
}
stages {
stage('Example') {
steps {
echo 'Hello World'
}
}
}
}
In agents
declared within a stage, the options are invoked before allocating the agent
and before checking any when
conditions.
In this case, when using timeout
, it is applied before the agent
is allocated.
The time to allocate the agent is included in the limit set by the timeout
option.
pipeline {
agent none
stages {
stage('Example') {
agent any
options {
// Timeout counter starts BEFORE agent is allocated
timeout(time: 1, unit: 'SECONDS')
}
steps {
echo 'Hello World'
}
}
}
}
This timeout will include the agent provisioning time. Because the timeout includes the agent provisioning time, the Pipeline may fail in cases where agent allocation is delayed.
In order to support the wide variety of use-cases Pipeline authors may have, the agent
section supports a few different types of parameters.
These parameters can be applied at the top-level of the pipeline
block, or within each stage
directive.
Execute the Pipeline, or stage, on any available agent.
For example: agent any
When applied at the top-level of the pipeline
block no global agent will be allocated for the entire Pipeline run and each stage
section will need to contain its own agent
section.
For example: agent none
Execute the Pipeline, or stage, on an agent available in the Jenkins environment with the provided label.
For example: agent { label 'my-defined-label' }
Label conditions can also be used:
For example: agent { label 'my-label1 && my-label2' }
or agent { label 'my-label1 || my-label2' }
agent { node { label 'labelName' } }
behaves the same as agent { label 'labelName' }
, but node
allows for additional options (such as customWorkspace
).
Execute the Pipeline, or stage, with the given container which will be dynamically provisioned on a node pre-configured to accept Docker-based Pipelines, or on a node matching the optionally defined label
parameter.
docker
also optionally accepts an args
parameter which may contain arguments to pass directly to a docker run
invocation, and an alwaysPull
option, which will force a docker pull
even if the image name is already present.
For example: agent { docker 'maven:3.9.3-eclipse-temurin-17' }
or
agent {
docker {
image 'maven:3.9.3-eclipse-temurin-17'
label 'my-defined-label'
args '-v /tmp:/tmp'
}
}
docker
also optionally accepts a registryUrl
and registryCredentialsId
parameters which will help to specify the Docker Registry to use and its credentials.
The parameter registryCredentialsId
could be used alone for private repositories within the docker hub.
For example:
agent {
docker {
image 'myregistry.com/node'
label 'my-defined-label'
registryUrl 'https://myregistry.com/'
registryCredentialsId 'myPredefinedCredentialsInJenkins'
}
}
Execute the Pipeline, or stage, with a container built from a Dockerfile
contained in the source repository.
In order to use this option, the Jenkinsfile
must be loaded from either a Multibranch Pipeline or a Pipeline from SCM.
Conventionally this is the Dockerfile
in the root of the source repository: agent { dockerfile true }
.
If building a Dockerfile
in another directory, use the dir
option: agent { dockerfile { dir 'someSubDir' } }
.
If your Dockerfile
has another name, you can specify the file name with the filename
option. You can pass additional arguments to the docker build …
command with the additionalBuildArgs
option, like agent { dockerfile { additionalBuildArgs '--build-arg foo=bar' } }
.
For example, a repository with the file build/Dockerfile.build
, expecting a build argument version
:
agent {
// Equivalent to "docker build -f Dockerfile.build --build-arg version=1.0.2 ./build/
dockerfile {
filename 'Dockerfile.build'
dir 'build'
label 'my-defined-label'
additionalBuildArgs '--build-arg version=1.0.2'
args '-v /tmp:/tmp'
}
}
dockerfile
also optionally accepts a registryUrl
and registryCredentialsId
parameters which will help to specify the Docker Registry to use and its credentials.
For example:
agent {
dockerfile {
filename 'Dockerfile.build'
dir 'build'
label 'my-defined-label'
registryUrl 'https://myregistry.com/'
registryCredentialsId 'myPredefinedCredentialsInJenkins'
}
}
Execute the Pipeline, or stage, inside a pod deployed on a Kubernetes cluster.
In order to use this option, the Jenkinsfile
must be loaded from either a Multibranch Pipeline or a Pipeline from SCM.
The Pod template is defined inside the kubernetes { } block.
For example, if you want a pod with a Kaniko container inside it, you would define it as follows:
agent {
kubernetes {
defaultContainer 'kaniko'
yaml '''
kind: Pod
spec:
containers:
- name: kaniko
image: gcr.io/kaniko-project/executor:debug
imagePullPolicy: Always
command:
- sleep
args:
- 99d
volumeMounts:
- name: aws-secret
mountPath: /root/.aws/
- name: docker-registry-config
mountPath: /kaniko/.docker
volumes:
- name: aws-secret
secret:
secretName: aws-secret
- name: docker-registry-config
configMap:
name: docker-registry-config
'''
}
You will need to create a secret aws-secret
for Kaniko to be able to authenticate with ECR.
This secret should contain the contents of ~/.aws/credentials
.
The other volume is a ConfigMap which should contain the endpoint of your ECR registry.
For example:
{
"credHelpers": {
"<your-aws-account-id>.dkr.ecr.eu-central-1.amazonaws.com": "ecr-login"
}
}
Refer to the following example for reference: https://github.com/jenkinsci/kubernetes-plugin/blob/master/examples/kaniko.groovy
These are a few options that can be applied to two or more agent
implementations.
They are not required unless explicitly stated.
A string.
The label or label condition on which to run the Pipeline or individual stage
.
This option is valid for node
, docker
, and dockerfile
, and is required for node
.
A string.
Run the Pipeline or individual stage
this agent
is applied to within this custom workspace, rather than the default.
It can be either a relative path, in which case the custom workspace will be under the workspace root on the node, or an absolute path.
For example:
agent {
node {
label 'my-defined-label'
customWorkspace '/some/other/path'
}
}
This option is valid for node
, docker
, and dockerfile
.
A boolean, false by default. If true, run the container on the node specified at the top-level of the Pipeline, in the same workspace, rather than on a new node entirely.
This option is valid for docker
and dockerfile
, and only has an effect when used on an agent
for an individual stage
.
A string.
Runtime arguments to pass to docker run
.
This option is valid for docker
and dockerfile
.
pipeline {
agent { docker 'maven:3.9.3-eclipse-temurin-17' } (1)
stages {
stage('Example Build') {
steps {
sh 'mvn -B clean verify'
}
}
}
}
1 | Execute all the steps defined in this Pipeline within a newly created container of the given name and tag (maven:3.9.3-eclipse-temurin-17 ). |
pipeline {
agent none (1)
stages {
stage('Example Build') {
agent { docker 'maven:3.9.3-eclipse-temurin-17' } (2)
steps {
echo 'Hello, Maven'
sh 'mvn --version'
}
}
stage('Example Test') {
agent { docker 'openjdk:17-jre' } (3)
steps {
echo 'Hello, JDK'
sh 'java -version'
}
}
}
}
1 | Defining agent none at the top-level of the Pipeline ensures that an Executor will not be assigned unnecessarily.
Using agent none also forces each stage section to contain its own agent section. |
2 | Execute the steps in this stage in a newly created container using this image. |
3 | Execute the steps in this stage in a newly created container using a different image from the previous stage. |
The post
section defines one or more additional steps that are run upon the completion of a Pipeline’s or stage’s run (depending on the location of the post
section within the Pipeline). post
can support any of the following post-condition blocks: always
, changed
, fixed
, regression
, aborted
, failure
, success
, unstable
, unsuccessful
, and cleanup
.
These condition blocks allow the execution of steps inside each condition depending on the completion status of the Pipeline or stage.
The condition blocks are executed in the order shown below.
Required |
No |
---|---|
Parameters |
None |
Allowed |
In the top-level |
always
Run the steps in the post
section regardless of the completion status of the Pipeline’s or stage’s run.
changed
Only run the steps in post
if the current Pipeline’s run has a different completion status from its previous run.
fixed
Only run the steps in post
if the current Pipeline’s run is successful and the previous run failed or was unstable.
regression
Only run the steps in post
if the current Pipeline’s or status is failure, unstable, or aborted and the previous run was successful.
aborted
Only run the steps in post
if the current Pipeline’s run has an "aborted" status, usually due to the Pipeline being manually aborted.
This is typically denoted by gray in the web UI.
failure
Only run the steps in post
if the current Pipeline’s or stage’s run has a "failed" status, typically denoted by red in the web UI.
success
Only run the steps in post
if the current Pipeline’s or stage’s run has a "success" status, typically denoted by blue or green in the web UI.
unstable
Only run the steps in post
if the current Pipeline’s run has an "unstable" status, usually caused by test failures, code violations, etc.
This is typically denoted by yellow in the web UI.
unsuccessful
Only run the steps in post
if the current Pipeline’s or stage’s run has not a "success" status.
This is typically denoted in the web UI depending on the status previously mentioned (for stages this may fire if the build itself is unstable).
cleanup
Run the steps in this post
condition after every other post
condition has been evaluated, regardless of the Pipeline or stage’s status.
pipeline {
agent any
stages {
stage('Example') {
steps {
echo 'Hello World'
}
}
}
post { (1)
always { (2)
echo 'I will always say Hello again!'
}
}
}
1 | Conventionally, the post section should be placed at the end of the Pipeline. |
2 | Post-condition blocks contain steps the same as the steps section. |
Containing a sequence of one or more stage directives, the stages
section is where the bulk of the "work" described by a Pipeline will be located.
At a minimum, it is recommended that stages
contain at least one stage directive for each discrete part of the continuous delivery process, such as Build, Test, and Deploy.
Required |
Yes |
---|---|
Parameters |
None |
Allowed |
Inside the |
pipeline {
agent any
stages { (1)
stage('Example') {
steps {
echo 'Hello World'
}
}
}
}
1 | The stages section will typically follow the directives such as agent , options , etc. |
The steps
section defines a series of one or more steps to be executed in a given stage
directive.
Required |
Yes |
---|---|
Parameters |
None |
Allowed |
Inside each |
pipeline {
agent any
stages {
stage('Example') {
steps { (1)
echo 'Hello World'
}
}
}
}
1 | The steps section must contain one or more steps. |
The environment
directive specifies a sequence of key-value pairs which will be defined as environment variables for all steps, or stage-specific steps, depending on where the environment
directive is located within the Pipeline.
This directive supports a special helper method credentials()
which can be used to access pre-defined Credentials by their identifier in the Jenkins environment.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Inside the |
The environment variable specified will be set to the Secret Text content.
The environment variable specified will be set to the location of the File file that is temporarily created.
The environment variable specified will be set to username:password
and two additional environment variables will be automatically defined: MYVARNAME_USR
and MYVARNAME_PSW
respectively.
The environment variable specified will be set to the location of the SSH key file that is temporarily created and two additional environment variables will be automatically defined: MYVARNAME_USR
and MYVARNAME_PSW
(holding the passphrase).
Unsupported credentials type causes the pipeline to fail with the message: |
pipeline {
agent any
environment { (1)
CC = 'clang'
}
stages {
stage('Example') {
environment { (2)
AN_ACCESS_KEY = credentials('my-predefined-secret-text') (3)
}
steps {
sh 'printenv'
}
}
}
}
1 | An environment directive used in the top-level pipeline block will apply to all steps within the Pipeline. |
2 | An environment directive defined within a stage will only apply the given environment variables to steps within the stage . |
3 | The environment block has a helper method credentials() defined which can be used to access pre-defined Credentials by their identifier in the Jenkins environment. |
pipeline {
agent any
stages {
stage('Example Username/Password') {
environment {
SERVICE_CREDS = credentials('my-predefined-username-password')
}
steps {
sh 'echo "Service user is $SERVICE_CREDS_USR"'
sh 'echo "Service password is $SERVICE_CREDS_PSW"'
sh 'curl -u $SERVICE_CREDS https://myservice.example.com'
}
}
stage('Example SSH Username with private key') {
environment {
SSH_CREDS = credentials('my-predefined-ssh-creds')
}
steps {
sh 'echo "SSH private key is located at $SSH_CREDS"'
sh 'echo "SSH user is $SSH_CREDS_USR"'
sh 'echo "SSH passphrase is $SSH_CREDS_PSW"'
}
}
}
}
The options
directive allows configuring Pipeline-specific options from within the Pipeline itself.
Pipeline provides a number of these options, such as buildDiscarder
, but they may also be provided by plugins, such as timestamps
.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Inside the |
Persist artifacts and console output for the specific number of recent Pipeline runs.
For example: options { buildDiscarder(logRotator(numToKeepStr: '1')) }
Perform the automatic source control checkout in a subdirectory of the workspace.
For example: options { checkoutToSubdirectory('foo') }
Disallow concurrent executions of the Pipeline.
Can be useful for preventing simultaneous accesses to shared resources, etc.
For example: options { disableConcurrentBuilds() }
to queue a build when there’s already an executing build of the Pipeline, or options { disableConcurrentBuilds(abortPrevious: true) }
to abort the running one and start the new build.
Do not allow the pipeline to resume if the controller restarts.
For example: options { disableResume() }
Used with docker
or dockerfile
top-level agent.
When specified, each stage will run in a new container instance on the same node, rather than all stages running in the same container instance.
Allows overriding default treatment of branch indexing triggers.
If branch indexing triggers are disabled at the multibranch or organization label, options { overrideIndexTriggers(true) }
will enable them for this job only.
Otherwise, options { overrideIndexTriggers(false) }
will disable branch indexing triggers for this job only.
Preserve stashes from completed builds, for use with stage restarting.
For example: options { preserveStashes() }
to preserve the stashes from the most recent completed build, or options { preserveStashes(buildCount: 5) }
to preserve the stashes from the five most recent completed builds.
Set the quiet period, in seconds, for the Pipeline, overriding the global default.
For example: options { quietPeriod(30) }
On failure, retry the entire Pipeline the specified number of times.
For example: options { retry(3) }
Skip checking out code from source control by default in the agent
directive.
For example: options { skipDefaultCheckout() }
Skip stages once the build status has gone to UNSTABLE.
For example: options { skipStagesAfterUnstable() }
Set a timeout period for the Pipeline run, after which Jenkins should abort the Pipeline.
For example: options { timeout(time: 1, unit: 'HOURS') }
pipeline {
agent any
options {
timeout(time: 1, unit: 'HOURS') (1)
}
stages {
stage('Example') {
steps {
echo 'Hello World'
}
}
}
}
1 | Specifying a global execution timeout of one hour, after which Jenkins will abort the Pipeline run. |
Prepend all console output generated by the Pipeline run with the time at which the line was emitted.
For example: options { timestamps() }
Set failfast true for all subsequent parallel stages in the pipeline.
For example: options { parallelsAlwaysFailFast() }
Completely disable option "Restart From Stage" visible in classic Jenkins UI and Blue Ocean as well.
For example: options { disableRestartFromStage() }
.
This option can not be used inside of the stage.
A comprehensive list of available options is pending the completion of help desk ticket 820. |
The options
directive for a stage
is similar to the options
directive at the root of the Pipeline.
However, the stage
-level options
can only contain steps like retry
, timeout
, or timestamps
, or Declarative options that are relevant to a stage
, like skipDefaultCheckout
.
Inside a stage
, the steps in the options
directive are invoked before entering the agent
or checking any when
conditions.
Skip checking out code from source control by default in the agent
directive.
For example: options { skipDefaultCheckout() }
Set a timeout period for this stage, after which Jenkins should abort the stage.
For example: options { timeout(time: 1, unit: 'HOURS') }
pipeline {
agent any
stages {
stage('Example') {
options {
timeout(time: 1, unit: 'HOURS') (1)
}
steps {
echo 'Hello World'
}
}
}
}
1 | Specifying an execution timeout of one hour for the Example stage, after which Jenkins will abort the Pipeline run. |
On failure, retry this stage the specified number of times.
For example: options { retry(3) }
Prepend all console output generated during this stage with the time at which the line was emitted.
For example: options { timestamps() }
The parameters
directive provides a list of parameters that a user should provide when triggering the Pipeline.
The values for these user-specified parameters are made available to Pipeline steps via the params
object, refer to the Parameters, Declarative Pipeline for its specific usage.
Each parameter has a Name and Value, depending on the parameter type.
This information is exported as environment variables when the build starts, allowing subsequent parts of the build configuration to access those values.
For example, use the ${PARAMETER_NAME}
syntax with POSIX shells like bash
and ksh
, the ${Env:PARAMETER_NAME}
syntax with PowerShell, or the %PARAMETER_NAME%
syntax with Windows cmd.exe
.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Only once, inside the |
A parameter of a string type, for example: parameters { string(name: 'DEPLOY_ENV', defaultValue: 'staging', description: '') }
.
A text parameter, which can contain multiple lines, for example: parameters { text(name: 'DEPLOY_TEXT', defaultValue: 'One\nTwo\nThree\n', description: '') }
.
A boolean parameter, for example: parameters { booleanParam(name: 'DEBUG_BUILD', defaultValue: true, description: '') }
.
A choice parameter, for example: parameters { choice(name: 'CHOICES', choices: ['one', 'two', 'three'], description: '') }
.
The first value is the default.
A password parameter, for example: parameters { password(name: 'PASSWORD', defaultValue: 'SECRET', description: 'A secret password') }
.
pipeline {
agent any
parameters {
string(name: 'PERSON', defaultValue: 'Mr Jenkins', description: 'Who should I say hello to?')
text(name: 'BIOGRAPHY', defaultValue: '', description: 'Enter some information about the person')
booleanParam(name: 'TOGGLE', defaultValue: true, description: 'Toggle this value')
choice(name: 'CHOICE', choices: ['One', 'Two', 'Three'], description: 'Pick something')
password(name: 'PASSWORD', defaultValue: 'SECRET', description: 'Enter a password')
}
stages {
stage('Example') {
steps {
echo "Hello ${params.PERSON}"
echo "Biography: ${params.BIOGRAPHY}"
echo "Toggle: ${params.TOGGLE}"
echo "Choice: ${params.CHOICE}"
echo "Password: ${params.PASSWORD}"
}
}
}
}
A comprehensive list of available parameters is pending the completion of help desk ticket 820. |
The triggers
directive defines the automated ways in which the Pipeline should be re-triggered.
For Pipelines which are integrated with a source such as GitHub or BitBucket, triggers
may not be necessary as webhooks-based integration will likely already be present.
The triggers currently available are cron
, pollSCM
and upstream
.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Only once, inside the |
Accepts a cron-style string to define a regular interval at which the Pipeline should be re-triggered, for example: triggers { cron('H */4 * * 1-5') }
.
Accepts a cron-style string to define a regular interval at which Jenkins should check for new source changes.
If new changes exist, the Pipeline will be re-triggered.
For example: triggers { pollSCM('H */4 * * 1-5') }
Accepts a comma-separated string of jobs and a threshold.
When any job in the string finishes with the minimum threshold, the Pipeline will be re-triggered.
For example: triggers { upstream(upstreamProjects: 'job1,job2', threshold: hudson.model.Result.SUCCESS) }
The |
// Declarative //
pipeline {
agent any
triggers {
cron('H */4 * * 1-5')
}
stages {
stage('Example') {
steps {
echo 'Hello World'
}
}
}
}
The Jenkins cron syntax follows the syntax of the cron utility (with minor differences). Specifically, each line consists of 5 fields separated by TAB or whitespace:
MINUTE | HOUR | DOM | MONTH | DOW |
---|---|---|---|---|
Minutes within the hour (0–59) |
The hour of the day (0–23) |
The day of the month (1–31) |
The month (1–12) |
The day of the week (0–7) where 0 and 7 are Sunday. |
To specify multiple values for one field, the following operators are available. In the order of precedence,
*
specifies all valid values
M-N
specifies a range of values
M-N/X
or */X
steps by intervals of X
through the specified range or whole valid range
A,B,…,Z
enumerates multiple values
To allow periodically scheduled tasks to produce even load on the system, the symbol H
(for “hash”) should be used wherever possible.
For example, using 0 0 * * *
for a dozen daily jobs will cause a large spike at midnight.
In contrast, using H H * * *
would still execute each job once a day, but not all at the same time, better using limited resources.
The H
symbol can be used with a range.
For example, H H(0-7) * * *
means some time between 12:00 AM (midnight) to 7:59 AM.
You can also use step intervals with H
, with or without ranges.
The H
symbol can be thought of as a random value over a range, but it actually is a hash of the job name, not a random function, so that the value remains stable for any given project.
Beware that for the day of month field, short cycles such as */3
or H/3
will not work consistently near the end of most months, due to variable month lengths.
For example, */3
will run on the 1st, 4th, …31st days of a long month, then again the next day of the next month.
Hashes are always chosen in the 1-28 range, so H/3
will produce a gap between runs of between 3 and 6 days at the end of a month.
Longer cycles will also have inconsistent lengths, but the effect may be relatively less noticeable.
Empty lines and lines that start with #
will be ignored as comments.
In addition, @yearly
, @annually
, @monthly
, @weekly
, @daily
, @midnight
, and @hourly
are supported as convenient aliases.
These use the hash system for automatic balancing.
For example, @hourly
is the same as H * * * *
and could mean at any time during the hour.
@midnight
actually means some time between 12:00 AM and 2:59 AM.
every fifteen minutes (perhaps at :07, :22, :37, :52) |
|
every ten minutes in the first half of every hour (three times, perhaps at :04, :14, :24) |
|
once every two hours at 45 minutes past the hour starting at 9:45 AM and finishing at 3:45 PM every weekday. |
|
once in every two hours slot between 9 AM and 5 PM every weekday (perhaps at 10:38 AM, 12:38 PM, 2:38 PM, 4:38 PM) |
|
once a day on the 1st and 15th of every month except December |
|
The stage
directive goes in the stages
section and should contain a steps section, an optional agent
section, or other stage-specific directives.
Practically speaking, all of the real work done by a Pipeline will be wrapped in one or more stage
directives.
Required |
At least one |
---|---|
Parameters |
One mandatory parameter, a string for the name of the stage. |
Allowed |
Inside the |
// Declarative //
pipeline {
agent any
stages {
stage('Example') {
steps {
echo 'Hello World'
}
}
}
}
A section defining tools to auto-install and put on the PATH
.
This is ignored if agent none
is specified.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Inside the |
The input
directive on a stage
allows you to prompt for input, using the input
step.
The stage
will pause after any options
have been applied, and before entering the agent
block for that stage
or evaluating the when
condition of the stage
.
If the input
is approved, the stage
will then continue.
Any parameters provided as part of the input
submission will be available in the environment for the rest of the stage
.
Required.
This will be presented to the user when they go to submit the input
.
An optional identifier for this input
.
The default value is based on the stage
name.
Optional text for the "ok" button on the input
form.
An optional comma-separated list of users or external group names who are allowed to submit this input
.
Defaults to allowing any user.
An optional name of an environment variable to set with the submitter
name, if present.
An optional list of parameters to prompt the submitter to provide. Refer to parameters for more information.
pipeline {
agent any
stages {
stage('Example') {
input {
message "Should we continue?"
ok "Yes, we should."
submitter "alice,bob"
parameters {
string(name: 'PERSON', defaultValue: 'Mr Jenkins', description: 'Who should I say hello to?')
}
}
steps {
echo "Hello, ${PERSON}, nice to meet you."
}
}
}
}
The when
directive allows the Pipeline to determine whether the stage should be executed depending on the given condition.
The when
directive must contain at least one condition.
If the when
directive contains more than one condition, all the child conditions must return true for the stage to execute.
This is the same as if the child conditions were nested in an allOf
condition (refer to the examples below).
If an anyOf
condition is used, note that the condition skips remaining tests as soon as the first "true" condition is found.
More complex conditional structures can be built using the nesting conditions: not
, allOf
, or anyOf
.
Nesting conditions may be nested to any arbitrary depth.
Required |
No |
---|---|
Parameters |
None |
Allowed |
Inside a |
Execute the stage when the branch being built matches the branch pattern (ANT style path glob) given, for example: when { branch 'master' }
. Note that this only works on a multibranch Pipeline.
The optional parameter comparator
may be added after an attribute to specify how any patterns are evaluated for a match:
EQUALS
for a simple string comparison
GLOB
(the default) for an ANT style path glob (same as for example changeset
)
REGEXP
for regular expression matching
For example: when { branch pattern: "release-\\d+", comparator: "REGEXP"}
Execute the stage when the build is building a tag.
For example: when { buildingTag() }
Execute the stage if the build’s SCM changelog contains a given regular expression pattern, for example: when { changelog '.*^\\[DEPENDENCY\\] .+$' }
.
Execute the stage if the build’s SCM changeset contains one or more files matching the given pattern.
Example: when { changeset "**/*.js" }
The optional parameter comparator
may be added after an attribute to specify how any patterns are evaluated for a match:
EQUALS
for a simple string comparison
GLOB
(the default) for an ANT style path glob case insensitive (this can be turned off with the caseSensitive
parameter).
REGEXP
for regular expression matching
For example: when { changeset pattern: ".TEST\\.java", comparator: "REGEXP" }
or when { changeset pattern: "*/*TEST.java", caseSensitive: true }
Executes the stage if the current build is for a "change request" (a.k.a. Pull Request on GitHub and Bitbucket, Merge Request on GitLab, Change in Gerrit, etc.).
When no parameters are passed the stage runs on every change request, for example: when { changeRequest() }
.
By adding a filter attribute with parameter to the change request, the stage can be made to run only on matching change requests.
Possible attributes are id
, target
, branch
, fork
, url
, title
, author
, authorDisplayName
, and authorEmail
.
Each of these corresponds to a CHANGE_*
environment variable, for example: when { changeRequest target: 'master' }
.
The optional parameter comparator
may be added after an attribute to specify how any patterns are evaluated for a match:
EQUALS
for a simple string comparison (the default)
GLOB
for an ANT style path glob (same as for example changeset
)
REGEXP
for regular expression matching
Example: when { changeRequest authorEmail: "[\\w_-.]+@example.com", comparator: 'REGEXP' }
Execute the stage when the specified environment variable is set to the given value, for example: when { environment name: 'DEPLOY_TO', value: 'production' }
.
Execute the stage when the expected value is equal to the actual value, for example: when { equals expected: 2, actual: currentBuild.number }
.
Execute the stage when the specified Groovy expression evaluates to true, for example: when { expression { return params.DEBUG_BUILD } }
.
When returning strings from your expressions they must be converted to booleans or return null to evaluate to false. Simply returning "0" or "false" will still evaluate to "true".
|
Execute the stage if the TAG_NAME
variable matches the given pattern.
For example: when { tag "release-*" }
If an empty pattern is provided the stage will execute if the TAG_NAME
variable exists (same as buildingTag()
).
The optional parameter comparator
may be added after an attribute to specify how any patterns are evaluated for a match:
EQUALS
for a simple string comparison,
GLOB
(the default) for an ANT style path glob (same as for example changeset
), or
REGEXP
for regular expression matching.
For example: when { tag pattern: "release-\\d+", comparator: "REGEXP"}
Execute the stage when the nested condition is false.
Must contain one condition.
For example: when { not { branch 'master' } }
Execute the stage when all of the nested conditions are true.
Must contain at least one condition.
For example: when { allOf { branch 'master'; environment name: 'DEPLOY_TO', value: 'production' } }
Execute the stage when at least one of the nested conditions is true.
Must contain at least one condition.
For example: when { anyOf { branch 'master'; branch 'staging' } }
Execute the stage when the current build has been triggered by the param given. For example:
when { triggeredBy 'SCMTrigger' }
when { triggeredBy 'TimerTrigger' }
when { triggeredBy 'BuildUpstreamCause' }
when { triggeredBy cause: "UserIdCause", detail: "vlinde" }
when
before entering agent
in a stage
By default, the when
condition for a stage
will be evaluated after entering the agent
for that stage
, if one is defined.
However, this can be changed by specifying the beforeAgent
option within the when
block.
If beforeAgent
is set to true
, the when
condition will be evaluated first, and the agent
will only be entered if the when
condition evaluates to true.
when
before the input
directiveBy default, the when condition for a stage will not be evaluated before the input, if one is defined.
However, this can be changed by specifying the beforeInput
option within the when block.
If beforeInput
is set to true, the when condition will be evaluated first, and the input will only be entered if the when condition evaluates to true.
beforeInput true
takes precedence over beforeAgent true
.
when
before the options
directiveBy default, the when
condition for a stage
will be evaluated after entering the options
for that stage
, if any are defined.
However, this can be changed by specifying the beforeOptions
option within the when
block.
If beforeOptions
is set to true
, the when
condition will be evaluated first, and the options
will only be entered if the when
condition evaluates to true.
beforeOptions true
takes precedence over beforeInput true
and beforeAgent true
.
pipeline {
agent any
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
branch 'production'
}
steps {
echo 'Deploying'
}
}
}
}
pipeline {
agent any
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
branch 'production'
environment name: 'DEPLOY_TO', value: 'production'
}
steps {
echo 'Deploying'
}
}
}
}
pipeline {
agent any
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
allOf {
branch 'production'
environment name: 'DEPLOY_TO', value: 'production'
}
}
steps {
echo 'Deploying'
}
}
}
}
pipeline {
agent any
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
branch 'production'
anyOf {
environment name: 'DEPLOY_TO', value: 'production'
environment name: 'DEPLOY_TO', value: 'staging'
}
}
steps {
echo 'Deploying'
}
}
}
}
pipeline {
agent any
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
expression { BRANCH_NAME ==~ /(production|staging)/ }
anyOf {
environment name: 'DEPLOY_TO', value: 'production'
environment name: 'DEPLOY_TO', value: 'staging'
}
}
steps {
echo 'Deploying'
}
}
}
}
beforeAgent
pipeline {
agent none
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
agent {
label "some-label"
}
when {
beforeAgent true
branch 'production'
}
steps {
echo 'Deploying'
}
}
}
}
beforeInput
pipeline {
agent none
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
beforeInput true
branch 'production'
}
input {
message "Deploy to production?"
id "simple-input"
}
steps {
echo 'Deploying'
}
}
}
}
beforeOptions
pipeline {
agent none
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
beforeOptions true
branch 'testing'
}
options {
lock label: 'testing-deploy-envs', quantity: 1, variable: 'deployEnv'
}
steps {
echo "Deploying to ${deployEnv}"
}
}
}
}
triggeredBy
pipeline {
agent none
stages {
stage('Example Build') {
steps {
echo 'Hello World'
}
}
stage('Example Deploy') {
when {
triggeredBy "TimerTrigger"
}
steps {
echo 'Deploying'
}
}
}
}
Stages in Declarative Pipeline may have a stages
section containing a list of nested stages to be run in sequential order.
A stage must have one and only one of steps , stages , parallel , or matrix .
It is not possible to nest a parallel or matrix block within a stage directive if that stage directive is nested within a parallel or matrix block itself.
However, a stage directive within a parallel or matrix block can use all other functionality of a stage , including agent , tools , when , etc.
|
pipeline {
agent none
stages {
stage('Non-Sequential Stage') {
agent {
label 'for-non-sequential'
}
steps {
echo "On Non-Sequential Stage"
}
}
stage('Sequential') {
agent {
label 'for-sequential'
}
environment {
FOR_SEQUENTIAL = "some-value"
}
stages {
stage('In Sequential 1') {
steps {
echo "In Sequential 1"
}
}
stage('In Sequential 2') {
steps {
echo "In Sequential 2"
}
}
stage('Parallel In Sequential') {
parallel {
stage('In Parallel 1') {
steps {
echo "In Parallel 1"
}
}
stage('In Parallel 2') {
steps {
echo "In Parallel 2"
}
}
}
}
}
}
}
}
Stages in Declarative Pipeline may have a parallel
section containing a list of nested stages to be run in parallel.
A stage must have one and only one of steps , stages , parallel , or matrix .
It is not possible to nest a parallel or matrix block within a stage directive if that stage directive is nested within a parallel or matrix block itself.
However, a stage directive within a parallel or matrix block can use all other functionality of a stage , including agent , tools , when , etc.
|
In addition, you can force your parallel
stages to all be aborted when any one of them fails, by adding failFast true
to the stage
containing the parallel
.
Another option for adding failfast
is adding an option to the pipeline definition: parallelsAlwaysFailFast()
.
pipeline {
agent any
stages {
stage('Non-Parallel Stage') {
steps {
echo 'This stage will be executed first.'
}
}
stage('Parallel Stage') {
when {
branch 'master'
}
failFast true
parallel {
stage('Branch A') {
agent {
label "for-branch-a"
}
steps {
echo "On Branch A"
}
}
stage('Branch B') {
agent {
label "for-branch-b"
}
steps {
echo "On Branch B"
}
}
stage('Branch C') {
agent {
label "for-branch-c"
}
stages {
stage('Nested 1') {
steps {
echo "In stage Nested 1 within Branch C"
}
}
stage('Nested 2') {
steps {
echo "In stage Nested 2 within Branch C"
}
}
}
}
}
}
}
}
parallelsAlwaysFailFast
pipeline {
agent any
options {
parallelsAlwaysFailFast()
}
stages {
stage('Non-Parallel Stage') {
steps {
echo 'This stage will be executed first.'
}
}
stage('Parallel Stage') {
when {
branch 'master'
}
parallel {
stage('Branch A') {
agent {
label "for-branch-a"
}
steps {
echo "On Branch A"
}
}
stage('Branch B') {
agent {
label "for-branch-b"
}
steps {
echo "On Branch B"
}
}
stage('Branch C') {
agent {
label "for-branch-c"
}
stages {
stage('Nested 1') {
steps {
echo "In stage Nested 1 within Branch C"
}
}
stage('Nested 2') {
steps {
echo "In stage Nested 2 within Branch C"
}
}
}
}
}
}
}
}
Stages in Declarative Pipeline may have a matrix
section defining a multi-dimensional matrix of name-value combinations to be run in parallel.
We’ll refer these combinations as "cells" in a matrix.
Each cell in a matrix can include one or more stages to be run sequentially using the configuration for that cell.
A stage must have one and only one of steps , stages , parallel , or matrix .
It is not possible to nest a parallel or matrix block within a stage directive if that stage directive is nested within a parallel or matrix block itself.
However, a stage directive within a parallel or matrix block can use all other functionality of a stage , including agent , tools , when , etc.
|
In addition, you can force your matrix
cells to all be aborted when any one of them fails, by adding failFast true
to the stage
containing the matrix
.
Another option for adding failfast
is adding an option to the pipeline definition: parallelsAlwaysFailFast()
.
The matrix
section must include an axes
section and a stages
section.
The axes
section defines the values for each axis
in the matrix.
The stages
section defines a list of stage
s to run sequentially in each cell.
A matrix
may have an excludes
section to remove invalid cells from the matrix.
Many of the directives available on stage
, including agent
, tools
, when
, etc., can also be added to matrix
to control the behavior of each cell.
The axes
section specifies one or more axis
directives.
Each axis
consists of a name
and a list of values
.
All the values from each axis are combined with the others to produce the cells.
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
}
// ...
}
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
axis {
name 'BROWSER'
values 'chrome', 'edge', 'firefox', 'safari'
}
}
// ...
}
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
axis {
name 'BROWSER'
values 'chrome', 'edge', 'firefox', 'safari'
}
axis {
name 'ARCHITECTURE'
values '32-bit', '64-bit'
}
}
// ...
}
The stages
section specifies one or more stage
s to be executed sequentially in each cell.
This section is identical to any other stages
section.
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
}
stages {
stage('build') {
// ...
}
stage('test') {
// ...
}
stage('deploy') {
// ...
}
}
}
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
axis {
name 'BROWSER'
values 'chrome', 'edge', 'firefox', 'safari'
}
}
stages {
stage('build-and-test') {
// ...
}
}
}
The optional excludes
section lets authors specify one or more exclude
filter expressions that select cells to be excluded from the expanded set of matrix cells (aka, sparsening).
Filters are constructed using a basic directive structure of one or more of exclude axis
directives each with a name
and values
list.
The axis
directives inside an exclude
generate a set of combinations (similar to generating the matrix cells).
The matrix cells that match all the values from an exclude
combination are removed from the matrix.
If more than one exclude
directive is supplied, each is evaluated separately to remove cells.
When dealing with a long list of values to exclude, exclude axis
directives can use notValues
instead of values
.
These will exclude cells that do not match one of the values passed to notValues
.
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
axis {
name 'BROWSER'
values 'chrome', 'edge', 'firefox', 'safari'
}
axis {
name 'ARCHITECTURE'
values '32-bit', '64-bit'
}
}
excludes {
exclude {
axis {
name 'PLATFORM'
values 'mac'
}
axis {
name 'ARCHITECTURE'
values '32-bit'
}
}
}
// ...
}
Exclude the linux, safari
combination and exclude any platform that is not windows
with the edge
browser.
matrix {
axes {
axis {
name 'PLATFORM'
values 'linux', 'mac', 'windows'
}
axis {
name 'BROWSER'
values 'chrome', 'edge', 'firefox', 'safari'
}
axis {
name 'ARCHITECTURE'
values '32-bit', '64-bit'
}
}
excludes {
exclude {
// 4 cells
axis {
name 'PLATFORM'
values 'mac'
}
axis {
name 'ARCHITECTURE'
values '32-bit'
}
}
exclude {
// 2 cells
axis {
name 'PLATFORM'
values 'linux'
}
axis {
name 'BROWSER'
values 'safari'
}
}
exclude {
// 3 more cells and '32-bit, mac' (already excluded)
axis {
name 'PLATFORM'
notValues 'windows'
}
axis {
name 'BROWSER'
values 'edge'
}
}
}
// ...
}
Matrix lets users efficiently configure the overall environment for each cell, by adding stage-level directives under matrix
itself.
These directives behave the same as they would on a stage but they can also accept values provided by the matrix for each cell.
The axis
and exclude
directives define the static set of cells that make up the matrix.
That set of combinations is generated before the start of the pipeline run.
The "per-cell" directives, on the other hand, are evaluated at runtime.
These directives include:
pipeline {
parameters {
choice(name: 'PLATFORM_FILTER', choices: ['all', 'linux', 'windows', 'mac'], description: 'Run on specific platform')
}
agent none
stages {
stage('BuildAndTest') {
matrix {
agent {
label "${PLATFORM}-agent"
}
when { anyOf {
expression { params.PLATFORM_FILTER == 'all' }
expression { params.PLATFORM_FILTER == env.PLATFORM }
} }
axes {
axis {
name 'PLATFORM'
values 'linux', 'windows', 'mac'
}
axis {
name 'BROWSER'
values 'firefox', 'chrome', 'safari', 'edge'
}
}
excludes {
exclude {
axis {
name 'PLATFORM'
values 'linux'
}
axis {
name 'BROWSER'
values 'safari'
}
}
exclude {
axis {
name 'PLATFORM'
notValues 'windows'
}
axis {
name 'BROWSER'
values 'edge'
}
}
}
stages {
stage('Build') {
steps {
echo "Do Build for ${PLATFORM} - ${BROWSER}"
}
}
stage('Test') {
steps {
echo "Do Test for ${PLATFORM} - ${BROWSER}"
}
}
}
}
}
}
}
Declarative Pipelines may use all the available steps documented in the Pipeline Steps reference, which contains a comprehensive list of steps, with the addition of the steps listed below which are only supported in Declarative Pipeline.
The script
step takes a block of Scripted Pipeline and executes that in the Declarative Pipeline.
For most use-cases, the script
step should be unnecessary in Declarative Pipelines, but it can provide a useful "escape hatch".
script
blocks of non-trivial size and/or complexity should be moved into Shared Libraries instead.
pipeline {
agent any
stages {
stage('Example') {
steps {
echo 'Hello World'
script {
def browsers = ['chrome', 'firefox']
for (int i = 0; i < browsers.size(); ++i) {
echo "Testing the ${browsers[i]} browser"
}
}
}
}
}
}
Scripted Pipeline, like Declarative Pipeline, is built on top of the underlying Pipeline sub-system. Unlike Declarative, Scripted Pipeline is effectively a general-purpose DSL [1] built with Groovy. Most functionality provided by the Groovy language is made available to users of Scripted Pipeline, which means it can be a very expressive and flexible tool with which one can author continuous delivery pipelines.
Scripted Pipeline is serially executed from the top of a Jenkinsfile
downwards, like most traditional scripts in Groovy or other languages.
Providing flow control, therefore, rests on Groovy expressions, such as the if/else
conditionals, for example:
if
, Scripted Pipelinenode {
stage('Example') {
if (env.BRANCH_NAME == 'master') {
echo 'I only execute on the master branch'
} else {
echo 'I execute elsewhere'
}
}
}
Another way Scripted Pipeline flow control can be managed is with Groovy’s exception handling support.
When Steps fail for whatever reason they throw an exception.
Handling behaviors on-error must make use of the try/catch/finally
blocks in Groovy, for example:
node {
stage('Example') {
try {
sh 'exit 1'
}
catch (exc) {
echo 'Something failed, I should sound the klaxons!'
throw
}
}
}
As discussed at the start of this chapter, the most fundamental part of a Pipeline is the "step". Fundamentally, steps tell Jenkins what to do and serve as the basic building block for both Declarative and Scripted Pipeline syntax.
Scripted Pipeline does not introduce any steps which are specific to its syntax; Pipeline Steps reference contains a comprehensive list of steps provided by Pipeline and plugins.
In order to provide durability, which means that running Pipelines can survive a restart of the Jenkins controller, Scripted Pipeline must serialize data back to the controller.
Due to this design requirement, some Groovy idioms such as collection.each { item → /* perform operation */ }
are not fully supported.
Refer to JENKINS-27421 and JENKINS-26481 for more information.
This video shares some differences between Scripted and Declarative Pipeline syntax.
When Jenkins Pipeline was first created, Groovy was selected as the foundation. Jenkins has long shipped with an embedded Groovy engine to provide advanced scripting capabilities for admins and users alike. Additionally, the implementors of Jenkins Pipeline found Groovy to be a solid foundation upon which to build what is now referred to as the "Scripted Pipeline" DSL. [1].
As it is a fully-featured programming environment, Scripted Pipeline offers a tremendous amount of flexibility and extensibility to Jenkins users. The Groovy learning-curve isn’t typically desirable for all members of a given team, so Declarative Pipeline was created to offer a simpler and more opinionated syntax for authoring Jenkins Pipeline.
Both are fundamentally the same Pipeline sub-system underneath. They are both durable implementations of "Pipeline as code". They are both able to use steps built into Pipeline or provided by plugins. Both are able to utilize Shared Libraries
Where they differ however is in syntax and flexibility. Declarative limits what is available to the user with a more strict and pre-defined structure, making it an ideal choice for simpler continuous delivery pipelines. Scripted provides very few limits, insofar that the only limits on structure and syntax tend to be defined by Groovy itself, rather than any Pipeline-specific systems, making it an ideal choice for power-users and those with more complex requirements. As the name implies, Declarative Pipeline encourages a declarative programming model. [2] Whereas Scripted Pipelines follow a more imperative programming model. [3]
Please submit your feedback about this page through this quick form.
Alternatively, if you don't wish to complete the quick form, you can simply indicate if you found this page helpful?
See existing feedback here.