The Infrastructure Platforms section enables GitLab Engineering to build and deliver safe, scalable and efficient features for multi-tenant and single-tenant GitLab SaaS platforms (GitLab.com and GitLab Dedicated).
Vision
To deliver on the mission, we are in the process of formalising the building blocks we need to work on.
We collaborate on section level items here. This channel is used to share important information with the wider team, but also serves to align all teams in Platfroms with the common topic.
Dedicated function channel. Please use this channel to ask questions about features or ways of using the Dedicated product. Dedicated group will use this channel to make announcements relevant to wider groups
Dedicated Switchboard team channel. Used to discuss topics that affect Switchboard team only. For broader engineering discussions please use #g_dedicated-team
Dedicated Environment Automation team channel. Used to discuss topics that affect Switchboard team only. For broader engineering discussions please use #g_dedicated-team
We use project specific channels to keep make it easier to follow specific topics. Channel names follow this format.
Queries to the Infrastructure team can be directed to a dedicated Slack channel #infrastructure_lounge. The Infrastructure department actively engages with these queries or Slack messages here, which are then converted into issues in a separate GitLab project tracker for efficient tracking and resolution.
Meetings
Once per week, we hold a Platforms leads call to align on action items related to career development, general direction or answer any ongoing questions that have not been addressed async. The call is cancelled when there are no topics added on the morning of the call.
In addition to the Platforms leads call, we have some recurring events and reminders that can be viewed in the SaaS Platforms Leadership Calendar. Please add this to your Calendars to stay up-to-date with the various events.
The Director of Product and Director of Infrastructure for SaaS Platforms conduct weekly progress reviews. During Platforms Grand Review, they review progress across all groups with the goals of being informed of progress made, learning about existing blockers, and supporting the team. The review is private streamed to the GitLab Unfiltered channel because the review covers confidential issues. All recordings are made available in the Platforms Grand Review YouTube Playlist.
Infrastructure Platforms Leads Demo
The Infrastructure Platforms Leads Demo is an opportunity for sync discussions between Staff+ IC across the Infrastructure Platforms Group to highlight current ongoing efforts underway in the teams they support.
All team members are welcome to join the call, but the emphasis is on Staff+ ICs to present and discuss the work they’re focused on, the problems they’re experiencing, and solutions they’re considering.
While the intention is for the call to be made public on GitLab Unfiltered, the default is for it to be published as private.
At the end of the call, a quick vote is held between the attendees and if all agree that the content is #SAFE, it can be published as public.
Slack to GitLab Issue Tracker Integration
In an effort to enhance the tracking and resolution of requests directed to the Infrastructure team, we are evaluating a bot that converts Slack messages in #infrastructure_lounge channel into GitLab issues.
Workflow Overview
Acknowledgement: An agent responds with the acknowledged_emoji (👀 in our case) to acknowledge a Slack message in the Infrastructure Lounge channel.
Issue Creation: The Slack bot then creates an issue with the acknowledging agent assigned to it.
Thread Attachment: The Slack thread corresponding to the message is also posted on the created GitLab issue.
Label Assignment: Agents can further categorize issues by adding label emojis (ops, foundations, scalability-observability or scalability-practices) in the Slack message. This action automatically assigns the issue to the respective team: Ops, Foundations, Scalability-Observability or Scalability-Practices.
Issue Closure: Agents/Requester can close the issue when resolved by adding any of the resolved_emojis (green-circle-check,white_check_markor checkedin our case)
Configuration
Agents responsible for handling these issues are defined in a JSON file, which serves as a CI/CD variable. Currently, this file contains a static list of all members of the infrastructure department.
Project and Backlog Management
We use epics and issues to manage our work. Our project management process is shared between all teams in SaaS Plaforms.
Tools
The Platforms section builds and maintains various tools to help deploy, operate and monitor our SaaS platforms. You can view a list of these tools in the Platforms Tools Index.
OKR
We use objective and key results to set goals in alignment with OKRs at GitLab.
Our OKR process is shared between all teams in Saas Platforms.
All team members are encouraged to schedule time for personal development. The following links may help you get started with Platforms-relevant learning. Please add your own contributions to this list to help others with their personal development.
Tools The Platforms section builds and maintains various tools to help deploy, operate and monitor our SaaS platforms. The below table is an index to help with the discovery and organization of tools that are actively maintained:
Tool Description Tamland Capacity planning forecasts for GitLab.com Stage Group Ownership Index Index of stage groups and their owned objects Stage Group Error Budgets Objective metrics to determine the reliability of a service Service Maturity Model Overview of each service’s operating capabilities Runway GitLab’s internal Platform as a Service implementation
All links under infrastructure/team/reliability/ have been moved or deprecated. If you are seeing this due to accessing a page under that path, we have probably deleted or relocated the content.
If you believe the content you are trying to access should still be in the handbook, please post a message in the #production_engineering Slack channel so we can see if the content is still relevant.
OKRs in SaaS Platforms Creating OKRs OKRs (or other items outside of projects) that require progress tracking should be updated every Wednesday.
When writing OKRs, the guidance is that:
Objective is defined as “What do you want to achieve?” Key Results is defined as “How will you know when you’ve achieved the objective?” As part of a KR, you can also have a sub point - which will likely tie to an epic.
Project Management in SaaS Platforms We use GitLab epics and issues to communicate the progress and status of our work. The SaaS Platforms epic is indexing the top level epic for each team, and links to active OKR’s for a given quarter. All teams in SaaS Platforms follow these guidelines so that it is easy for team-members to contribute to different projects if needed.
Projects are reviewed weekly in the Grand Review Every Wednesday, the DRI for a project is expected to update the status block in the epic description to:
