Ticketing Style Guide

Why do we need a Ticketing Style Guide?

To maintain the quality of our customer experience as the Support team grows, we have gathered a collection of best practices and suggestions for styling and responding to Zendesk tickets.

Consistency is what enables any Support Engineer to jump into an unfamiliar ticket and get to work fairly quickly. Consistency enables scaling, and consistency is what we’re hoping to achieve with this guide.

Some of the suggestions in this guide will encounter cases that warrant exceptions, and that’s OK.

Everyone in the Support Team is welcome to contribute or modify the suggestions in this guide given that they provide the “Why”, especially when a practice has proven to work, or not work for them through experience.

Readability

1. If your reply consists of more than 6 lines, consider breaking it to sections using HEADERS. Example:

### Logs analysis

In reviewing your logs we've noticed the following highlights:
  1. ...
  2. ...
  3. ...

### Next steps

It looks like you're hitting [this issue](issue-link)

Let's try the following:
  1. Perform some action `A` followed by a `gitlab-ctl reconfigure`.
  2. If the output is `X`, then try `Y`.

This will help the customer as well as fellow support engineers to quickly find the information they’re looking for.

1. Aim to add a public 1 line summary of the latest current issue for every 3 replies you send to the customer. Example:

Thanks for providing all this info.

### Current Status

So far we've established that you're facing a 500 error intermittently when loading an MR. This also seems to happen only for MRs with large diff content.

### More sections

This will ensure that the customer and you remain on the same page regarding what the issue is.

This will also help other support engineers to jump into the ticket and catch up on the context more efficiently.

Technical Content

1. Ask for gitlab.rb, or any other missing information or logs, but don’t let your reply consist of only that. Describe why you’re asking for the additional info. Example:

Would you please provide your `gitlab.rb` for all your nodes? Specifically, I'm interested in seeing your Redis configuration for each node.

It is possible that you've configured the caching differently in each node, which could result in inconsistent behaviors such as what we're seeing in this case.

This will help the customer understand how they can debug this issue themselves.

It will also help other support engineers to understand what you were aiming to look for in the provided data, and keep the debugging direction consistent over the next replies.

1. When requesting logs from the customer, only ask for as much as you think you’ll need to diagnose the problem. Example:

If the issue is reproducible and is not front-end related, ask for `gitlab-ctl tail`, if it's intermittent, ask for GitlabSOS.

This makes it easier to debug the issue as you’ll go through the minimal number of log files, and save the customer from breaking up a large GitLabSOS file, unless GitlabSOS is necessary.

View page source - Edit this page - please contribute.