Automations

What are Zendesk automations?

As per Zendesk:

Automations are similar to triggers because both define conditions and actions that modify ticket properties and optionally send email notifications to customers and agents. Where they differ is that automations execute when a time event occurs after a ticket property was set or updated, rather than immediately after a ticket is created or updated.

The simpler way to think of it is automations are triggers that do not run instantly. They are time based than event based.

Change management

Keep in mind, all change management should be stemming from an issue, first and foremost.

Creating a new automation with managed content

When your new automation is going to be using managed content, you will first need to get the managed content file in the Support managed content project. Remember to use the correct filenames for all of this to prevent Pipeline error “No managed content file” in the sync repo project later on.

Only after that has been done should you proceed to the next steps, which will match the steps detailed in Creating a new automation without managed content exactly.

Creating a new automation without managed content

As we manage our automations via sync repos, you simply need to create the automation file within the sync repo project. From there, the sync will create the automation itself within Zendesk.

Updating an existing automation

Updating an existing automation is considerably easier than creating a new one. Simply change the code in the source project and it will occur via the sync repo.

The one caveat you need to consider is when you are changing an automation to allow for managed content (or to disable it using managed content).

If you are adding managed content for the automation, see Creating a new automation with managed content as that process will detail setting up the connection.

If you are removing managed content for the automation, you will simply change the automation file in the source sync repo project via your merge request. After that has been merged, you will want to comment on the original issue asking the requester to remove the file from the corresponding Support managed content project.

Deactivating an automation

To deactivate an automation, you will simply change the automation file in the source sync repo project via your merge request. Ensure you merge request does the following:

• Moves the file from the data/active folder to the data/inactive folder
• Sets active: true to active: false in the file.
• Sets all conditions of:
  • Ticket: Type is not Incident
  • Ticket: Status category less than Solved
• Set an action of:
  • Ticket: Type Incident
  • Ticket: Add tags XXXX (replacing XXXX with the automation’s ID)

Deleting a deactivated automation

NOTE We avoid doing this unless an automation has been deactivated for a full year. After that point it can be deleted completely. Do also note that this will result in a complete change to positions and can cause the need for subsequent merge requests to the sync repo project.

To delete an automation, you need to purge it from multiple locations:

• Sync repo project
• Support managed content project
• Zendesk itself

The first two can be done via merge requests, but the last one has to be done in the Zendesk instance itself. To do this, open up the admin page of your corresponding Zendesk instance (Global or US Government), click Objects and rules on the left-hand side, and then click Automations. On this page, you will want to click Inactive, hover over the automation you are deleting, click the three vertical dots at right-hand side of automation, and click Delete. This will cause a pop-up modal to appear asking you to confirm the action. Click blue Delete automation button to do so.

Troubleshooting

Pipeline error “No managed content file”

This happens when we have said a managed content file should exist, but the git submodule does not contain one. This is commonly caused by:

• The file does not actually exist. If this is the case, you need to assist in getting it created in the Support managed content project
• Filename mismatches. This all works very specifically using naming conventions. If there is something even slightly off, your pipelines will encounter issues. The scripts are looking for a file that has the exact same name as the automation’s title. So if your automation has a title of Jason::Do a Thing, the corresponding Support managed content project should have a file with the same name (Jason::Do a Thing.md for managed content, Jason::Do a Thing.webhook for webhooks, or Jason::Do a Thing.email for emails). You will need to assist in correcting that on the Support managed content project first, and then rebase your merge request after that is done.
• You created the merge request in the source project before the file was added to the Support managed content project. To rectify this, get the Support managed content project MR completed and merged first. Once that has been done, you can rebase your MR by making a comment of /rebase. After it performs the rebase, your MR’s CI/CD pipeline should pass.

Pipeline error “Blank ID”

This means the script detected a YAML file within data/active or data/inactive that has an id value of blank (or nil). You will need to locate the file mentioned in the error and correct that.

Pipeline error “Blank position”

This means the script detected a YAML file within data/active or data/inactive that has an position value of blank (or nil). You will need to locate the file mentioned in the error and correct that.

Pipeline error “Blank title”

This means the script detected a YAML file within data/active or data/inactive that has an title value of blank (or nil). You will need to locate the file mentioned in the error and correct that.

Pipeline error “Inactive automation in active folder”

This means the script detected a YAML file within data/active that has an active value of false. You will need to locate the file mentioned in the error and correct that.

Pipeline error “Active automation in inactive folder”

This means the script detected a YAML file within data/inactive that has an active value of anything other than false. You will need to locate the file mentioned in the error and correct that.

Pipeline error “GitLab errors”

This is a generic error message that will detail some error that occurred when trying to either create or update the tag used on the source project. The exact steps to fix this will vary based on the nature of the error itself. You will need to review the error and determine the next steps from there.

If you are unsure how to proceed, it is best to seek assistance from the wider team.

Source Projects

Zendesk Global

• Support managed content project
• Sync repo project

Zendesk US Government

• Support managed content project
• Sync repo project

Positioning

Many components of Zendesk using positioning to determine the overall run order. With automations being time-based events, it is not often as important to worry about positioning as with something like triggers or views. We should still strive to be very deliberate in the positioning we use. A good thought to have is “what order would I want these running in if they all ran at once?”

By default, new automations gain a position of N+1, where N is the highest position value of all automations currently in Zendesk (both active and inactive). This is desired and we should rarely need to change this.

To edit positions in the Zendesk UI, go to the automations page. From there, click the three horizontal dots at the top-right of the page (on the same line as the search bar). That will bring up a sub-menu with the option Reorder page. Clicking that will then allow you to drag and drop the list of automations into the order you desire. After making the changes, click the blue Save button at the top right of the page.

Note: Both active and inactive automations have a integer positional value. While this does not matter in the UI, it will matter in the repo sync we utilize.

Automation standards

To ensure all automations we utilize are both consistent in nature and transparent in their actions, we strive to meet some standards on all automations we work with.

Naming standards

As Zendesk automations do not support categorization at this time, we have implemented a naming standard to help categorize the automations we have. This standard is as follows:

What it impacts::What it does/Who it impacts::Name of automation

Example 1

If you were making an automation to send a notification to Jason once a ticket has been in an open state for more than 24 hours, you would use the automation name of:

Notifications::Jason::Notify ticket has been open for more than 24 hours

This is because:

• the What it impacts is “Notifications”, since it sends a notification.
• the What it does/Who it impacts is “Jason”, since it sends a notification to Jason.
• the Name of automation can be whatever the creator wants for this, but it should be relatively short and describe the automation in a way that anyone who knows our naming standards can look at it and know what it does.

Example 2

If you were making an automation that sets a ticket to Closed after it has been in the state of Solved for 24 hours, you would use the automation name of:

Status::Close::Autoclose after 24 hours solved

This is because:

• the What it impacts is “Status”, since it impacts a ticket’s status.
• the What it does/Who it impacts is “Close”, since it closes a ticket.
• the Name of automation can be whatever the creator wants for this, but it should be relatively short and describe the automation in a way that anyone who knows our naming standards can look at it and know what it does.

Condition standards

Generally speaking, we aim to make automation conditions as simple as possible. When possible, you should use condition sets that are very specific and succinct. As an example, if you wanted an automation to only run when the form is Support Ops, it is better to simply put a condition of “Form is Support Ops” than adding exclusions for every other form. This can take time and practice to learn, so when in doubt, pair with the rest of the Support Ops team!