What is Charity Majors’s stance on SRE?#
Reviewed 28 September 2026. This report interprets Majors’s public writing from 2016 through 2026; her descriptions of team types are views about suitable organizational choices, not a universal definition of SRE.
Current finding#
Charity Majors values SRE’s engineering practices and the operational skill behind them, but rejects treating Google’s SRE organization as a template for every company. [1] [2] She argues that reliable services require developers to own the code they run, supported where needed by specialists in operations, SRE, and platform engineering. [3] [4] Her recurring objection is to assigning reliability to a renamed team while leaving operational work, authority, and incentives unchanged. [2] [5]
What she values in SRE#
In her 2016 response to the “DevOps vs SRE” debate, Majors called Google’s SRE book valuable for scaling systems and organizations and praised its hard technical and cultural work. [1] She also credited the SRE movement, alongside DevOps, with helping establish operations engineers as engineers rather than less valued system administrators. [2]
Majors uses SRE practices in her own account of production excellence. She recommends agreed service level objectives (SLOs) to align engineering and management on acceptable reliability, protect time to address service risks, and guide responses to an error budget. [6] She pairs those objectives with observability, shared expertise, blameless incident work, and structural improvements. [6] This overlaps with the broader SRE approach to service objectives described in the Phase 1 overview . [7]
In March 2026, Majors described herself among SREs and emphasized their focus on outcomes such as whether a service stays up. [8] Discussing AI-generated software, she argued that SREs should bring their experience building guardrails to new development workflows. [8]
Why context matters more than the label#
Majors’s 2016 criticism was directed at applying Google’s model wholesale. She argued that Google’s scale and operating context make some of its practices powerful there but potentially harmful if copied prescriptively into a different organization. [1] She regarded DevOps as a broader, more adaptable philosophy, while recognizing that large scale or difficult operational problems can justify specialized operations engineering. [1]
In 2022, Majors described SRE as most familiar in very large companies where software teams generally own their production code, with SREs often embedded in teams or product groups. [4] In her description, such specialists concentrate on reliability, instrumentation, observability, incident response, blameless reviews, coordination, and scaling. [4] She distinguished this from platform teams, whose principal customer is the developer and whose job is to make service ownership easier through self-service tools and sensible defaults; she allowed for an adjacent SRE team where reliability work warrants one. [4]
Ownership without erasing operations#
Majors has long argued that operational outcomes are produced across an organization and cannot be handed entirely to a separate ops team. [1] [2] Application engineers should be connected to what their code does in production, including through on-call work, instrumentation, and debugging. [3] Specialization remains useful, but she warns against compartmentalization that separates developers from those outcomes. [1]
Her January 2026 essay sharpened the other side of that argument: asking developers to own their services does not mean asking each team to own the entire infrastructure stack. [5] She describes product development and operations as complementary concerns, with operations providing a resilient, scalable, efficient base for products. [5] She argues that calling that function SRE, DevOps, or platform engineering does not itself improve operational outcomes, and that “ops” should be recognized as skilled engineering rather than treated as a synonym for toil. [5] Her view leaves room for a dedicated operations or SRE team when an organization actually has hard operational problems, while requiring clarity about which responsibilities that team takes on. [4] [5]
Where she differs from Google’s SRE methodology#
The clearest differences concern how to organize and describe the work, rather than whether reliability needs engineering, measurement, and collaboration. [1] [6] [9]
Where to start?#
Google’s original model starts with a software-engineering-led operations team that builds software to replace manual work and reserves at least half its time for engineering projects. [10]
Majors starts with the organization’s actual operational needs; she supports specialist teams when scale or hard operational problems justify them, rather than copying Google’s staffing model. [1] [5]
Who runs the service?#
Google says developers own services by default, but its engagement model says SRE typically handles most operational work after a supported service reaches general availability, while developers retain some incident and on-call exposure. [9] [11]
Majors puts stronger emphasis on developers running and debugging the code they build, with platform and operations specialists making that feasible without requiring every developer to manage the entire infrastructure stack. [3] [4] [5]
What is “ops”?#
Google defines SRE as an engineering approach to operations and caps operational load so SREs can eliminate repetitive manual work. [10]
Majors stresses that operations itself is a valuable engineering function, not a synonym for toil, and that changing its team label cannot repair poor ownership or incentives. [2] [5]
How broad is the frame?#
Google’s SRE framework is deliberately service-oriented and prescribes tools such as SLOs and error budgets for managing reliability. [9]
Majors looks more broadly at organizational culture, operational skill, and division of responsibility; she argues that the wider DevOps philosophy adapts more readily to differing contexts. [1] [2] [5]
These are differences of emphasis, not a claim that Google assigns all reliability to SRE or believes operations is only toil. Google’s later workbook explicitly says product developers own services by default and that SRE work is shared with them. [9] Its editors also acknowledged criticism that the first book was too centered on Google and cautioned readers against treating either book as a universal prescription. [12] That narrows the gap with Majors’s 2016 critique, though her 2026 defense of operations as a named, respected function remains a distinct emphasis. [1] [5] [12]
The stance in practice#
The practical reading of Majors’s position is to start with the service and the work: agree on user-relevant reliability targets, make the engineers who build a service accountable for its production behavior, and provide specialized support where scale or complexity demands it. [1] [3] [4] [6] Keep operational expertise visible and valued, whatever the team’s title. [2] [5] This is an interpretation of her writing, not a staffing formula she says every organization must follow. [1] [4]
An open question for a later pass is where, in a particular organization, service-team ownership should end and a dedicated SRE or operations team’s authority should begin. Majors describes several possible arrangements but does not prescribe one universal boundary. [1] [4] [5]