# Runner

Available in PagerDuty Process Automation products

The Runner, available for both Process Automation and Runbook Automation securely opens up network/communication between data centers and the Automation Cluster. The Runner is a Remote Execution hub for Node Steps to run on specified endpoints, rather than from the Automation server itself.

# Architecture

The Runner is a Java based program which uses a polling model to pick up work from the Automation Server. During each polling cycle (every 5 seconds) the Runner checks for executions that it is responsible for. Communication from the Runner to the Automation Server happens over https and is initiated from the Runner. This allows for enhanced firewall security as ports no longer need to be open for the Automation Server to talk to nodes over more sensitive ports. (e.g. SSH/22)

Runner Architecture
Runner Architecture

# Installation

Follow these steps to install a Runner:

# Runner Management

# Assigning Projects/Nodes to a Runner

Runners will execute Node Steps on Nodes within a project based on a Node Filter assigned to the Runner. To assign Projects and specific nodes to a Runner follow these steps:

  1. In the Runner Management screen click on the name of the Runner instance to be edited.
  2. If the project is not listed already choose Add Projects
  3. Toggle the Projects that should be added.
  4. The Node Filter section is used to identify Nodes that this runner is responsible for. By default .* will apply it to all nodes in the project.
  5. Click Add and the projects are now listed under the Runner configuration.

It is possible to edit each project Node Filter individually if needed. Use the Actions menu to Edit the project entry's Node Filter.

# Troubleshooting

Runner Logs are located in the ./runner/logs folder under the folder where the jar was executed from. The runner.log file contains operational and important messages about the runner. operations.log tracks an operation starts and if it succeeds or fails. Read more about logging and setting up custom logging.

# Proxying Runner connections

Runners can be configured to connect through a HTTP/HTTPS proxy. Proxies are commonly used to centralize and secure outbound traffic from the datacenter to internet services. The proxy configuration is optional and is added as java command line arguments when the runner process is started.

# Proxy configuration without proxy authentication

The following example will allow the runner to connect through the secure company proxy with address wp.acme.corp.

java -Dmicronaut.http.client.proxy-type=http -Dmicronaut.http.client.proxy-address=wp.acme.corp:443 -jar pdrunner.jar
  1. -Dmicronaut.http.client.proxy-type is set to http
  2. -Dmicronaut.http.client.proxy-address is set to the secure proxy company address.

# Proxy configuration with proxy authentication

The following example adds basic auth proxy configuration to the runner. The proxy-type and proxy-address settings are the same as the unauthenticated access example.

java -Dmicronaut.http.client.proxy-type=http -Dmicronaut.http.client.proxy-address=wp.acme.corp:443 -Dmicronaut.http.client.proxy-username=proxyUsernameString -Dmicronaut.http.client.proxy-password=proxyPassString -jar pdrunner.jar
  1. -Dmicronaut.http.client.proxy-username is set to the user that is allowed to connect through the secure proxy.
  2. -Dmicronaut.http.client.proxy-password is set to the secure proxy user password.


Does this replace Clustering?

This feature is intended to be used in situations where execution of Node Steps is needed in a network segment or security zone that is different from the Cluster feature(s). An Automation Server Cluster still provides high-availability and fault tolerance for the User Interface, Job Management/Execution and system management tasks.

Does this work for all communication into the remote data center?

At this time, the Runner will execute Node Executor and File Copier steps. This includes node steps such as Remote Command, Inline Script, Script File, and Copy File, as well as commands executed from the Commands tab. Node Sources, Health Checks, and other communication are not included in this version of the Runner.

Do Runners participate in node discovery?

Not at this time. To dynamically update nodes use custom scripting and the APIs to update the node lists remotely.

Can multiple Runners run in parallel?

Yes, multiple runners can be configured for the same project. If multiple runners have the same node scope, they race to retrieve the relevant task list. Multiple runners can be used to increase scale or add more redundancy in the automation architecture.