Dedicated Commercial (was Environment Automation)

Summary

Dedicated Commercial is a group of functional teams within the Dedicated Group. Our mission is to develop and operate the automated plumbing of the GitLab Dedicated solution, focused mainly on Dedicated Commercial roadmap. This group of functional teams is a spinoff from Environment Automation Team, after Act 2 at GitLab.

We follow the same processes as listed on the Dedicated Group, unless a difference exists which is explicitly noted on this page.

Team Members by Manager

Name Role
Oriol LluchOriol Lluch Manager, Infrastructure
Andrey RuzmanovAndrey Ruzmanov Senior Backend Engineer
Andy KnightAndy Knight Staff Site Reliability Engineer, Environment Automation
Dylan LauDylan Lau Site Reliability Engineer
Graeme GilliesGraeme Gillies Senior Site Reliability Engineer, Dedicated:Switchboard
Kunal ParmarKunal Parmar Senior Site Reliability Engineer
Muhamed HuseinbašićMuhamed Huseinbašić Site Reliability Engineer, Environment Automation
Site Reliability Engineer, Environment AutomationSite Reliability Engineer, Environment Automation Site Reliability Engineer, Environment Automation
Riccardo TrivellatoRiccardo Trivellato Site Reliability Engineer, Environment Automation
Samir HafezSamir Hafez Senior Site Reliability Engineer
Nick SkoretzNick Skoretz Site Reliability Engineer, Environment Automation
Tania RoblotTania Roblot Senior Site Reliability Engineer, Environment Automation
Vihang KotharkarVihang Kotharkar Site Reliability Engineer
Name Role
Stephen DenhamStephen Denham Manager, Dedicated:Environment Automation
Alexei OzerovaAlexei Ozerova Senior Site Reliability Engineer, Environment Automation
Chris TaylorChris Taylor Senior Site Reliability Engineer
Dave DittmanDave Dittman Site Reliability Engineer
Harpratap SinghHarpratap Singh Site Reliability Engineer
Jonathan BoweJonathan Bowe Site Reliability Engineer, Environment Automation
Konst TchernovKonst Tchernov Senior Site Reliability Engineer
Ryan PhamRyan Pham Site Reliability Engineer
Stephan BreitrainerStephan Breitrainer Senior Site Reliability Engineer, Environment Automation
Sean ArnoldSean Arnold Staff Backend Engineer, Dedicated:Switchboard
Veronica MondoVeronica Mondo Senior Site Reliability Engineer, Environment Automation

Functional Teams after Act 2

R&D department shifted to an organization structure that privileges smaller teams, 3-4 engineers, fully focused on a functional area, with a manager owning 2-3 teams. It also encouraged the role of Functional Lead, a Tech Lead for each functional team, Staff+ responsible for team technical execution and growth.

To fulfill this intent, Environment Automation was split into 5 functional teams:

Functional team Serves Scope and ownership Key alignment Manager Functional Lead
Platform Integration GitLab module development teams Owns the integration path that brings external products and modular GitLab features into Dedicated, including Dedicated binding, productionization, and the testing strategy for integration paths. Theseus and the Test Automation Framework. Interim: o-olluch@o-olluch Brendan McKitrickBrendan McKitrick
Platform Resilience Dedicated operations teams Owns resilience and self-healing products for individual tenants — including disaster recovery, autoscaling, automated capacity planning, rate limiting, and other mechanisms that allow an instance to detect, absorb, and recover from failures with minimal human intervention, while keeping infrastructure cost-effective. Dedicated Orchestration Automation and Cloud Cost Utilization. Stephen DenhamStephen Denham Konst TchernovKonst Tchernov
Dedicated Responder Automation Dedicated SRE and DevOps responders Owns responder workflows and tooling for Dedicated incidents, including signal and alert enrichment, alert routing, alert aggregation and automated troubleshooting. Incident Management and Signals/Observability in Production Engineering. o-olluch@o-olluch Tania RoblotTania Roblot
Dedicated Orchestration Automation Dedicated operators and DevOps teams Owns the core orchestration platform and change-management automation: job workflows, underlying infrastructure changes, maintenance operations, and tenant-model segments. Platform Resilience, Responder Automation, and Platform Integration consume the platform and own their workflows and configuration within it. Switchboard Orchestration APIs; also coordinates with Service Delivery Platform and Cells operations where interfaces overlap. o-olluch@o-olluch Samir HafezSamir Hafez
Dedicated Commercial Features Dedicated Product and Dedicated customers Owns the commercial Dedicated feature roadmap and production-ready implementations, including customer-facing infrastructure and isolation capabilities such as BYOD, network configuration, and Geo Proxy. Platform Integration and Dedicated Orchestration Automation for artifact injection, integration, and production quality. Stephen DenhamStephen Denham Veronica MondoVeronica Mondo

Working with us

We are in the transition to functional teams, so to engage with any functional team please refer to Environment Automation team (we will update this section when new communication channels are available per functional team):

How We Work

Our preference is to work asynchronously, within our project issue tracker as described in the project management section.

The team also has a set of regular synchronous calls:

  1. Environment Automation Team Sync (alternate weeks):
    1. EMEA/AMER: Tue 15:00 UTC (Good for EMEA and US East)
    2. PST/APAC: Wed 00:00 UTC (Good for APAC and US West)
  2. Dedicated on GCP - Weekly Demo: Wed 07:30 UTC
  3. Dedicated Group Demo

Quarterly Planning Process

Quarterly Planning Process

Quarterly planning is used to balance strategic priorities with team capacity. It ensures consistent execution, async alignment, and visibility across stakeholders. Planning starts 4 weeks before the new quarter and includes integrated Epic backlog refinement to ensure execution-ready work items.

Step 1 – Planning Issue Creation (Week –4)

  • Owner: Engineering Manager (EM)
  • Actions:
    • Create a Quarterly Planning Issue (example: Q4 FY26 Planning).
    • Use this issue for async discussion, links to strategic docs, and final decision logging.
    • Share issue link with PMs, team, and stakeholders.
  • Deliverables: Planning issue opened and shared with all stakeholders.

Step 2 – Strategic Input (Weeks –4 to –3)

  • Owners: PMs, Company Interlock process
  • Actions:
  • Deliverables: Strategic priorities documented in the planning issue.

Step 3 – Team Planning (Weeks –3 to –1)

  • Owners: EMs and Team Members
  • Actions:
    • EMs update the EA Continuous Planning Spreadsheet with projects, timeline estimates, and staffing needs.
    • Team members review estimates, flag risks, and surface dependencies.
    • EMs and team use comments in the planning issue to gauge capacity and connect their inputs.
  • Deliverables: Draft quarterly execution plan linked in the planning issue.

Step 4 – Epic Backlog Refinement (Weeks –2 to –1)

  • Owners: EM + PM with Team Members
  • Actions:
    • Take Epics identified for the upcoming quarter from the roadmap planning spreadsheet
    • For each Epic targeted for execution, ensure it contains:
      • MVC Scope clearly defined
      • Business Case / Rationale documented
      • Link to high-level design included
      • Estimated level of complexity assessed with engineering input
    • Move Epics through status progression: Triage → Proposal → Ready
    • Use planning issue comments for Epic refinement discussions and decisions
    • Team members provide technical input on complexity estimates and design feasibility
    • Pull in different stakeholders as necessary to complete Epic information
  • Deliverables: Refined Epic backlog with all targeted Epics in “Ready” status for immediate quarter execution.

Step 5 – Finalization (Week 0)

  • Owners: EMs with PM + stakeholders
  • Actions:
    • Async review in the planning issue.
    • Confirm scope, capacity allocation, and commitments.
    • Verify Epic backlog refinement completion and readiness status.
    • Post summary comment with final decisions, artifacts, and links.
    • EA EMs and PM use the weekly Tuesday conversation to review outstanding issues, flags, and confirm quarterly decisions.
  • Deliverables: Locked-in quarterly plan with refined Epic backlog; planning issue closed.

Step 6 – Quarterly Kickoff (Week 1)

  • Owners: EMs with PM + stakeholders
  • Actions:
    • PM and EMs collaborate on Dedicated Stage Kickoff Presentation (example: FY26Q3 Dedicated Kickoff)
    • PM and EMs fill in previous quarter achievements and next quarter initiatives
    • PMs and EMs record kickoff video and share with the team
    • Engineers can immediately begin work on refined, “Ready” status Epics
  • Deliverables: Quarterly kickoff presentation and recorded video for team alignment

Planning Inputs & Sources

  • Strategic: Company Interlock Process, PM Roadmap, Customer Commitments
  • Operational: KTLO Work, On-call Capacity, Unplanned Work, Engineering Backlog
  • Epic-Level: Technical designs, complexity assessments, MVC definitions

Planning Artifacts & Templates

Roles & Responsibilities

Role Responsibilities
Product Managers Maintain strategic roadmap, update planning spreadsheet, provide Epic business cases
Engineering Managers Lead capacity assessment, update EA Continuous Planning, create/manage planning issue, coordinate Epic refinement
Team Members Review estimates, flag risks, surface dependencies, provide Epic complexity input, async discussion
Stakeholders Provide feedback, validate commitments during Finalization, contribute to Epic refinement as needed

Success Metrics

  • Planning issue closed by Week 0
  • Team acknowledges capacity commitments
  • Strategic priorities mapped to execution capacity
  • Dependencies and risks clearly identified
  • All targeted Epics in “Ready” status with complete refinement information
  • Balanced allocation: Feature Dev 40–60%, KTLO 20–30%, On-call 15–25%, Engineering 10–15%

Risks & Best Practices

  • Assume 10–20% buffer for reactive/unplanned work
  • Document discussions inside the planning issue
  • Link spreadsheets and roadmaps for traceability
  • Review past quarter’s unplanned work
  • Balance strategic work with operational requirements
  • Complete Epic refinement before quarter start to enable immediate engineering execution
  • Ensure Epic complexity estimates include engineering perspective

Backlog Refinement Integration

This quarterly planning process includes integrated Epic backlog refinement during Step 4 (Weeks -2 to -1). The refinement ensures that engineers can quickly get started on an Epic once it’s ready to be picked up during the quarter execution phase.

Having this set of refined epics helps us plan for the upcoming quarter with accurate capacity estimates and allows engineers to immediately begin productive work when the quarter begins, rather than spending time on discovery and scoping activities.

Reviewer roulette

Reviewer roulette is an internal tool for use on GitLab.com projects that randomly picks a maintainer + reviewer. Environment Automation uses it to spread the MR review workload. To do so:

  1. Go to the reviewer roulette page.
  2. Click on Spin the wheel.

See the full MR process.

Epic Status Updates

Each week, automation will prompt DRIs for progress updates on their epics; these are collated and collectively reviewed by the business. DRIs should ensure their Epics are configured correctly and updates written or delegated in case of PTO.

Example responses

Here are some concrete examples of responses to capacity planning alerts.

  • Removing a metric from capacity planning - Advanced search memory pressure does not follow long-term trends and was not a useful prediction. It remains a metric that is alerted on if it exceeds practical limits.
  • Remove saturation metric entirely - kube_pool_cpu was incorrect in many cases, and difficult to get right. It needed to be replaced with a different saturation metric (node-based CPU).
  • Add Saturation metrics - Kubernetes PVCs were not being monitored at all, leading to near-miss incidents
  • Fix the saturation metric - Advanced search disk was inaccurate and needed to be replaced with better promql expressions
Last modified August 12, 2026: Added functional teams (d0e35fe1)