API Reference | Version 6

Rundeck provides a Web API for use with your application.

API Version Number

The current API version is 5.

For API endpoints described in this document, the minimum API version required for their use is indicated by the URL used, e.g.:

/api/2/system/info
/api/1/projects

This means you must use at least the API version indicated to access the endpoint, unless otherwise noted. Some features or functionality for the endpoint may only be supported in later versions.

The API Version Number is required to be included in all API calls within the URL.

If the version number is not included or if the requested version number is unsupported, then the API call will fail. The error response will include the code "api-version-unsupported" and have HTTP status code of 400 Bad Request:

<result error="true" apiversion="2">
    <error code="api-version-unsupported">
        <message>
        Unsupported API Version "1". API Request: /rundeck/api/1/project/test/jobs. Reason: Minimum supported version: 2
        </message>
    </error>
</result>

Changes

Changes introduced by API Version number:

Version 5:

Version 4:

Version 3:

Version 2:

URLs

The Rundeck server has a "Base URL", where you access the server. Your Rundeck Server URL may look like: http://myserver:4440.

The root URL path for all calls to the API in this version is:

$RUNDECK_SERVER_URL/api/2

XML

The majority of API calls require XML for input and output. Some import/export features support YAML formatted documents, but XML is used for all API-level information. (JSON support is planned for a future API version.)

Authentication

Authentication can be done in two different ways, either with Token based authentication, or with a username and password.

Note that in either case, it is recommended that you enable SSL Support for the Rundeck server so that communication is encrypted at all times. For more information about using SSL, see Administration - Configuring Rundeck for SSL.

Token Authentication

Token Authentication consists of including a string known as an "API Token" with every request to the Rundeck API.

To obtain an API Token, you must first log in to the Rundeck GUI using a user account that has "admin" credentials. Click on your username in the header of the page, and you will be shown your User Profile page. From this page you can manage your API Tokens. Click "Generate API Token" to create a new one. The unique string that is shown is the API Token.

You must include one of the following with every HTTP request to the API:

OR

With that in place, you can make calls to the API as described in the rest of this document, and you don't need to maintain any cookies between requests.

Examples:

Using the URL parameter to request the project list:

GET /api/1/projects?authtoken=E4rNvVRV378knO9dp3d73O0cs1kd0kCd HTTP/1.1
...

Using the HTTP Header:

GET /api/1/projects HTTP/1.1
X-Rundeck-Auth-Token: E4rNvVRV378knO9dp3d73O0cs1kd0kCd
...

Password Authentication

If using Password Authentication, you must perform the authentication steps prior to accessing the API.

This means that you must submit authentication parameters (username, password) to the "Authentication URL" and retain a Session Cookie.

The Session Cookie must be sent with all calls to the API to maintain the authenticated connection.

To submit authentication, submit a POST request to the URL:

$RUNDECK_SERVER_URL/j_security_check

With these parameters:

If the response includes a redirect chain which includes or results in $RUNDECK_SERVER_URL/user/login or $RUNDECK_SERVER_URL/user/error, then the authentication request failed.

Otherwise, if the response is a redirect chain which results in 200 successful response, then the authentication was successful.

The response should set a cookie named JSESSIONID.

Response Format

The XML Response format will conform to this document structure:

<result success/error="true" apiversion="X">
    <!-- error included if error="true" -->
    <error>
        <message><!-- error message text --></message>
        <!-- ... multiple message elements -->
    </error>

    <!-- optional success element if declared for the endpoint -->
    <success>
        <message><!-- success message --></message>
    </success>

    <!-- 
        Specific API results..
    -->
</result>

If an error occurred, then the error attribute of the <result> element will be "true". Otherwise a success attribute will have the value "true".

Some <error> responses will include a code attribute giving a specific type of error code, in addition to the message.

The apiversion attribute will be set to the latest version of the API supported by the server.

Error codes

Defined error codes that may be returned as <error code="...">

api-version-unsupported

The specified API version is not supported for the requested resource

unauthorized

The requested action is not authorized and/or the connection is not authenticated.

Item Lists

Many API requests will return an item list as a result. These are typically in the form:

<[items] count="x">
    <[item] ...>
    <[item] ...>
</[items]>

Where the list of specific items are wrapped in a pluralized element name which contains a count attribute.

When an API path declares its results as an "Item List" this is the format that will be returned.

API Contents

System Info

Get Rundeck server information and stats.

URL:

/api/1/system/info

Parameters: none

Result: Success response, with included system info and stats in this format:

<system>
    <timestamp epoch="1305909785806" unit="ms">
        <datetime>2011-05-20T16:43:05Z</datetime>
    </timestamp>
    <rundeck>
        <version>1.2.1</version>
        <apiversion>2</apiversion>
        <build>1.2.1-0-beta</build>
        <node>Venkman.local</node>
        <base>/Users/greg/rundeck121</base>
    </rundeck>
    <os>
        <arch>x86_64</arch>
        <name>Mac OS X</name>
        <version>10.6.7</version>
    </os>
    <jvm>
        <name>Java HotSpot(TM) 64-Bit Server VM</name>
        <vendor>Apple Inc.</vendor>
        <version>19.1-b02-334</version>
    </jvm>
    <stats>
        <uptime duration="300584" unit="ms">
            <since epoch="1305909485222" unit="ms">
                <datetime>2011-05-20T16:38:05Z</datetime>
            </since>
        </uptime>
        <cpu>
            <loadAverage unit="percent">0.40234375</loadAverage>
            <processors>4</processors>
        </cpu>
        <memory unit="byte">
            <max>477233152</max>
            <free>76626216</free>
            <total>257163264</total>
        </memory>
        <scheduler>
            <running>0</running>
        </scheduler>
        <threads>
            <active>24</active>
        </threads>
    </stats>
</system>

Description of included elements:

timestamp describes the current system time known to the server. The @epoch attribute includes the milliseconds since the unix epoch.

datetime
The W3C date and time

rundeck includes information about the Rundeck application.

rundeck/version

Rundeck version

rundeck/apiversion

Rundeck API version

rundeck/build

Rundeck build stamp

rundeck/node

Server node name

rundeck/base

Server base directory

os/arch

Operating System architecture

os/name

Operating System Name

os/version

Operating System Version

jvm/name

JVM name

jvm/vendor

JVM vendor

jvm/version

JVM version

stats section includes some system statistics:

uptime describes the JVM uptime as duration in ms, and includes absolute startup time:

uptime/since

JVM startup time as time since the unix epoch

uptime/since/datetime

JVM startup time as W3C date time.

cpu/loadAverage

JVM load average percentage for the system for the previous minute (see getSystemLoadAverage)

cpu/processors

Number of available system processors. note that loadAverage might be calculated based on the total number of available processors

The memory section describes memory usage in bytes:

max

Maximum JVM memory that can be allocated

free

Free memory of the allocated memory

total

Total allocated memory for the JVM

scheduler/running

Number of running jobs in the scheduler

threads/active

Number of active Threads in the JVM

Listing Jobs

List the jobs that exist for a project.

URL:

/api/1/jobs

Required parameters:

The following parameters can also be used to narrow down the result set.

Result: An Item List of jobs. Each job is of the form:

<job id="ID">
    <name>Job Name</name>
    <group>Job Name</group>
    <project>Project Name</project>
    <description>...</description>
</job>

Note: If neither groupPath nor groupPathExact are specified, then the default groupPath value of "*" will be used (matching jobs in all groups). groupPathExact cannot be combined with groupPath. You can set either one to "-" to match only the top-level jobs which are not within a group.

Listing Jobs for a Project

List the jobs that exist for a project. (API version 2 required.)

URL:

/api/2/project/[NAME]/jobs

The following parameters can also be used to narrow down the result set.

Result: An Item List of jobs. Each job is of the form:

<job id="ID">
    <name>Job Name</name>
    <group>Job Name</group>
    <project>Project Name</project>
    <description>...</description>
</job>

Note: If neither groupPath nor groupPathExact are specified, then the default groupPath value of "*" will be used (matching jobs in all groups). groupPathExact cannot be combined with groupPath. You can set either one to "-" to match only the top-level jobs which are not within a group.

Running a Job

Run a job specified by ID.

URL:

/api/1/job/[ID]/run

Optional parameters:

Result: An Item List of executions containing a single entry for the execution that was created. See Listing Running Executions.

Exporting Jobs

Export the job definitions for in XML or YAML formats.

URL:

/api/1/jobs/export

Required parameters:

Optional parameters:

The following parameters can also be used to narrow down the result set.

Result:

If you specify format=xml, then the output will be in jobs-v20(5) format.

If you specify format=yaml, then the output will be in jobs-v20-yaml(5) format.

If an error occurs, then the output will be in XML format, using the common result element described in the Response Format section.

Importing Jobs

Import job definitions in XML or YAML formats.

URL:

/api/1/jobs/import

Method: POST

Expected Content-Type: application/x-www-form-urlencoded (since 1.3) or multipart/form-data

Required Content:

Optional parameters:

Result:

A set of status results. Each imported job definition will be either "succeeded", "failed" or "skipped". These status sections contain a count attribute declaring how many jobs they contain. Within each one there will be 0 or more job elements.

<succeeded count="x">
    <!-- job elements -->
</succeeded>
<failed count="x">
    <!-- job elements -->
</failed>
<skipped count="x">
    <!-- job elements -->
</skipped>

Each Job element will be of the form:

<job>
    <!-- ID may not exist if the job was not created yet -->
    <id>ID</id>
    <name>job name</name>
    <group>job group</group>
    <project>job project</project>
    <!--if within the failed section, then an error section will be included -->
    <error>Error message</error> 
</job>

Getting a Job Definition

Export a single job definition in XML or YAML formats.

URL:

/api/1/job/[ID]

Optional parameters:

Result:

If you specify format=xml, then the output will be in jobs-v20(5) format.

If you specify format=yaml, then the output will be in jobs-v20-yaml(5) format.

If an error occurs, then the output will be in XML format, using the common result element described in the Response Format section.

Deleting a Job Definition

Delete a single job definition.

URL:

/api/1/job/[ID]

Method: DELETE

Result:

The common result element described in the Response Format section, indicating success or failure and any messages.

If successful, then the result will contain a success/message element with the result message:

<success>
<message>Job was successfully deleted: ... </message>
</success>

Bulk Job Delete

Delete multiple job definitions at once.

URL:

/api/5/jobs/delete

Method: DELETE, or POST

Query parameters:

Note: you can combine ids with idlist

Result:

The common result element described in the Response Format section, indicating success or failure and any messages.

If successful, then the result will contain a deleteJobs element with two sections of results, succeeded and failed:

<deleteJobs requestCount="#" allsuccessful="true/false">
    <succeeded count="1">
        <deleteJobRequest id="[job ID]">
            <message>[message]</message>
        </deleteJobRequest>
    </succeeded>
    <failed count="1">
        <deleteJobRequest id="[job ID]" errorCode="[code]">
            <error>[message]</error>
        </deleteJobRequest>
    </failed>
</deleteJobs>

deleteJobs will have two attributes:

The response may contain only one of succeeded or failed, depending on the result.

The succeeded and failed sections contain multiple deleteJobRequest elements.

Each deleteJobRequest under the succeeded section will contain:

Each deleteJobRequest under the failed section will contain:

Getting Executions for a Job

Get the list of executions for a Job.

URL:

/api/1/job/[ID]/executions

Optional Query Parameters:

Result: an Item List of executions. See Listing Running Executions.

Listing Running Executions

List the currently running executions for a project

URL:

/api/1/executions/running

Required Parameters:

Result: An Item List of executions. Each execution of the form:

<execution id="[ID]" href="[url]" status="[status]">
    <user>[user]</user>
    <date-started unixtime="[unixtime]">[datetime]</date-started>

    <!-- optional job context if the execution is associated with a job -->
    <job id="jobID" averageDuration="[milliseconds]">
        <name>..</name>
        <group>..</group>
        <description>..</description>
    </job>

    <!-- description of the execution -->
    <description>...</description>

    <!-- argString (arguments) of the execution -->
    <argstring>...</argstring>

    <!-- The following elements included only if the execution has ended -->

    <!-- the completion time of the execution -->
    <date-ended unixtime="[unixtime]">[datetime]</date-ended> 

    <!-- if the execution was aborted, the username who aborted it: -->
    <abortedby>[username]</abortedby>

</execution>

The [status] value indicates the execution status. It is one of:

The [url] value is a URL to the Rundeck server page to view the execution output.

[user] is the username of the user who started the execution.

[unixtime] is the millisecond unix timestamp, and [datetime] is a W3C dateTime string in the format "yyyy-MM-ddTHH:mm:ssZ".

If known, the average duration of the associated Job will be indicated (in milliseconds) as averageDuration. (Since API v5)

Getting Execution Info

Get the status for an execution by ID.

URL:

/api/1/execution/[ID]

Result: an Item List of executions with a single item. See Listing Running Executions.

Execution Query

Query for Executions based on Job or Execution details.

URL:

/api/5/executions

Result: an Item List of executions. See Listing Running Executions.

The executions element will have paging attributes:

Required Parameters:

The following parameters can also be used to narrow down the result set.

Parameters for Execution details:

The format for the end, and begin filters is either: a unix millisecond timestamp, or a W3C dateTime string in the format "yyyy-MM-ddTHH:mm:ssZ".

Parameters for querying for Executions for particular jobs:

The format for the jobListFilter and excludeJobListFilter is the job's group and name separated by a '/' character, such as: "group1/job name", or "my job" if there is no group.

Paging parameters:

Execution Output

Get the output for an execution by ID. The execution can be currently running or may have already completed.

URL:

/api/5/execution/[ID]/output

The log output for each execution is stored in a file on the Rundeck server, and this API endpoint allows you to retrieve some or all of the output, in several possible formats: json, XML, and plain text. When retrieving the plain text output, some metadata about the log is included in HTTP Headers. JSON and XML output formats include metadata about each output log line, as well as metadata about the state of the execution and log file, and your current index location in the file.

Several parameters can be used to retrieve only part of the output log data. You can use these parameters to more efficiently retrieve the log content over time while an execution is running.

The log file used to store the execution output is a formatted text file which also contains metadata about each line of log output emitted during an execution. Several data values in this API endpoint refer to "bytes", but these do not reflect the size of the final log data; they are only relative to the formatted log file itself. You can treat these byte values as opaque locations in the log file, but you should not try to correlate them to the actual textual log lines.

Optional Parameters:

Result: The output content in the requested format.

Tailing Output

To "tail" the output from a running execution, you will need to make a series of requests to this API endpoint, and update the offset value that you send to reflect the returned dataoffset value that you receive. This gives you a consistent pointer into the output log file.

When starting these requests, there are two mechanisms you can use:

  1. Start at the beginning, specifying either a lastmod or a offset of 0
  2. Start at the end, by using lastlines to receive the last available set of log lines.

After your first request you will have the dataoffset and lastmod response values you can use to continue making requests for subsequent log output. You can choose several ways to do this:

  1. Use the offset and lastmod parameters to indicate modification time and receive as much output as is available
  2. Use the offset and maxlines parameter to specify a maximum number of log entries
  3. Use only the offset parameter and receive as much output as is available.

After each request, you will update your offset value to reflect the dataoffset in the response.

All log output has been read when the iscompleted value is "true".

Below is some example pseudo-code for using this API endpoint to follow the output of a running execution "live":

Authorization:

This endpoint requires that the user have 'read' access to the Job or to Adhoc executions to retrieve the output content.

Output Format Using the URL

Specifying an output format can occur in several ways. The simplest ways are to include the format in the URL, either by including a format URL parameter, or an extension on the request URL.

When using a URL format, use one of these values for the format:

To use a URL parameter, add a ?format= parameter to your request.

E.g.:

GET /api/5/execution/3/output?format=json

To use a URL extension, add a ".[format]" to the end of the URL, but prior to any URL parameters.

E.g.:

GET /api/5/execution/3/output.xml?offset=120

Output Format using Accept Header

You can also specify the format using Content Negotiation techniques by including an Accept header in your request, and specifying a valid MIME-type to represent one of the formats:

E.g.:

GET /api/5/execution/3/output
Accept: */xml

Output Content

The result will contain a set of data values reflecting the execution's status, as well as the status and read location in the output file.

Entries:

Each log entry will be included in a section called entries.

Content of each Log Entry:

The XML entry will have the log message as the text value.

Text Format Content

For the plain text format, the content of the response will simply be the log output lines at the chosen offset location.

Included in the response will be some HTTP headers that provide the metadata about the output location. Some headers may not be present, depending on the state of the response. See the Output Content section for descriptions of the content and availability of the values:

Aborting Executions

Abort a running execution by ID.

URL:

/api/1/execution/[ID]/abort

Result: The result will contain a success/message element will contain a descriptive message. The status of the abort action will be included as an element:

<abort status="[abort-state]">
    <execution id="[id]" status="[status]"/>
</abort>

The [abort-state] will be one of: "pending", "failed", or "aborted".

Running Adhoc Commands

Run a command string.

URL:

/api/1/run/command

Required Parameters:

Optional Parameters:

Node filter parameters as described under Using Node Filters

Result: A success message, and a single <execution> item identifying the new execution by ID:

<execution id="X"/>

Running Adhoc Scripts

Run a script.

URL:

/api/1/run/script

Method: POST

Expected Content-Type: application/x-www-form-urlencoded

Required Parameters:

Required Content:

Optional Parameters:

Node filter parameters as described under Using Node Filters

Result: A success message, and a single <execution> item identifying the new execution by ID:

<execution id="X"/>

Running Adhoc Script URLs

Run a script downloaded from a URL. (API version 4 required.)

URL:

/api/4/run/url

Method: POST

Expected Content-Type: application/x-www-form-urlencoded

Required Parameters:

Required Content:

Optional Parameters:

Node filter parameters as described under Using Node Filters

Result: A success message, and a single <execution> item identifying the new execution by ID:

<execution id="X"/>

Listing Projects

List the existing projects on the server.

URL:

/api/1/projects

Result: An Item List of projects, each project of the form specified in the Getting Project Info section.

Getting Project Info

Get information about a project.

URL:

/api/1/project/[NAME]

Result: An Item List of projects with one project. The project is of the form:

<project>
    <name>Project Name</name>
    <description>...</description>
    <!-- additional items -->
</project>

If the project defines a Resource Model Provider URL, then the additional items are:

<resources>
    <providerURL>URL</providerURL>
</resources>

Updating and Listing Resources for a Project

Update or retrieve the Resources for a project. A GET request returns the resources for the project, and a POST request will update the resources. (API version 2 required.)

URL:

/api/2/project/[NAME]/resources

Method: POST, GET

POST request

POSTing to this URL will set the resources for the project to the content of the request.

Expected POST Content: For API version 2: either text/xml or text/yaml Content-Type containing the Resource Model definition in resources-v10(5) or resources-v10-yaml(5) formats as the request body. (Note: any MIME type ending with '/yaml' or '/x-yaml' or '/xml' will be accepted).

Since API version 3: You can also POST data using a content type supported by a Resource Format Parser plugin. This requires using API version 3.

POST Result: A success or failure API response. (See Response Format).

Example POST request:

POST /api/2/project/test/resources
Content-Type: text/yaml

node1:
  hostname: node1
  username: bob

node2:
  hostname: node2
  username: bob

Result:

200 OK

<result success="true">
    <success>
        <message>Resources were successfully updated for project test</message>
    </success>
</result>

GET request

Optional GET Parameters:

GET Result: Depending on the format parameter, a value of "xml" will return resources-v10(5) and "yaml" will return resources-v10-yaml(5) formatted results.

Example GET request:

GET /api/2/project/test/resources

Response:

200 OK
Content-Type: text/xml

<project>
    <node name="node1" hostname="node1" username="bob" />
    <node name="node2" hostname="node2" username="bob" />
</project>

Refreshing Resources for a Project

Refresh the resources for a project via its Resource Model Provider URL. The URL can be specified as a request parameter, or the pre-configured URL for the project can be used. (API version 2 required.)

URL:

/api/2/project/[NAME]/resources/refresh

Method: POST

Optional Parameters:

providerURL
Specify the Resource Model Provider URL to refresh the resources from. If not specified then the configured provider URL in the project.properties file will be used.

Result: A success or failure result with a message.

The URL requested as the providerURL must be allowed by the project.properties and framework.properties configuration settings according to these rules:

Multiple regexes can be specified in those config files by adding multiple properties:

project.resources.allowedURL.0=^http://myserver:9090/resources/.*$
project.resources.allowedURL.1=^http://server2:9090/resources/.*$

Listing History

List the event history for a project.

URL:

/api/1/history

Required Parameters:

Optional Parameters:

The format for the end, and begin filters is either: a unix millisecond timestamp, or a W3C dateTime string in the format "yyyy-MM-ddTHH:mm:ssZ".

The format for the jobListFilter and excludeJobListFilter is the job's group and name separated by a '/' character, such as: "group1/job name", or "my job" if there is no group.

Result: an Item List of events. Each event has this form:

<event starttime="[unixtime]" endtime="[unixtime]">
  <title>[job title, or "adhoc"]</title>
  <status>[status]</status>
  <summary>[summary text]</summary>
  <node-summary succeeded="[X]" failed="[Y]" total="[Z]"/>
  <user>[user]</user>
  <project>[project]</project>
  <date-started>[start date]</date-started>
  <date-ended>[end date]</date-ended>
  <!-- if the execution was aborted, the username who aborted it: -->
  <abortedby>[username]</abortedby>
  <!-- if associated with an Execution, include the execution id: -->
  <execution id="[execid]"/>
  <!-- if associated with a Job, include the Job ID: -->
  <job id="[jobid]"/>
</event>

The events element will also have max, offset, and total attributes, to indicate the state of paged results. E.G:

<events count="8" total="268" max="20" offset="260">
...
</events>

total is the total number of events matching the query parameters. count is the number of events included in the results. max is the paging size as specified in the request, or with the default value of 20. offset is the offset specified, or default value of 0.

Creating History Event Reports

Create a history event report for any external process.

URL:

/api/1/report/create

Required Parameters:

Optional Parameters:

The format for the end, and start values is either: a unix millisecond timestamp, or a W3C dateTime string in the format "yyyy-MM-ddTHH:mm:ssZ".

Listing Resources

List or query the resources for a project.

URL:

/api/1/resources

Required Parameters:

Optional Parameters:

Note: If no query parameters are included, the result set will include all Node resources for the project.

Result: Depending on the format parameter, a value of "xml" will return resources-v10(5) and "yaml" will return resources-v10-yaml(5) formatted results. Any other supported format value will return content in the specified format.

Using Node Filters

To include certain resources, specify the inclusion filters:

To exclude certain resources, specify the exclusion filters as above but with exclude- prepended:

To specify which type of filter (inclusion or exclusion) takes precedence, specify:

If using only inclusion filters, the result set will be only those resources which do match the filters.

If using only exclusion filters, the result set will be only those resources which do not match the filters.

When using both types of filters, the result will depend on the exclude-precedence value (default true).

The difference between these results is apparent when you have resources which are matched by both the exclusion and the inclusion filters. The precedence determines whether those resources are included or not.

Using set logic, if "A" is the set of all resources, "X" is the set of all resources matching the exclusion filters, and "I" is the set of all resources matching the inclusion filters, then:

Getting Resource Info

Get a specific resource within a project.

URL:

/api/1/resource/[NAME]

Required Parameters:

Optional Parameters:

Result: Depending on the format parameter, a value of "xml" will return resources-v10(5) and "yaml" will return resources-v10-yaml(5) formatted results.

The result will contain a single item for the specified resource.