What is site reliability engineering (SRE)?
Site reliability engineering is the process of utilizing operation processes and assigning them to the software engineering team for the automation.
Credit for the term “SRE” goes to Google’s Ben Treynor Sloss
Ben Treynor Sloss of Google is the mastermind behind SRE, and aptly describes it as “what happens when a software engineer is tasked with what used to be called operations”. The concept arose after an examination of the conflicts between operations, who want to ensure that features don’t break anything or inconvenience end users, and dev teams, who have developed and want to release new features as soon as they are ready for a rollout. SRE is a reconciliation between the two.
A team of Google engineers literally wrote the book on SRE
Google published a book on SRE that is available for free online. It offers a deep dive into the role of SRE and recommended best practices for execution. Parts II and III, principles and practices (respectively) are of note.
SRE Principles: The core principles of SRE, according to Google, are:
- Embracing risk: Provide neutral approaches to service management using error budgets.
- Service level objectives: Provides recommendations for disintengled indicators from agreements and examines how SRE uses the terms.
- Eliminating toil: Stepping away from mundane and repetitive tasks that are devoid of value.
- Monitoring distributed systems: Always avoid being blind to what is going on in the organization for the sake of reliability.
- Release engineering: Carefully account for releases to ensure that they are consistent and do not contribute to outages.
- Simplicity: A system that is too complex can lower reliability and become difficult to scale back to a simpler place.
SRE Practices: SREs run related systems for external or internal users, and are responsible for the services. Successful operation of the services include: capacity planning, addressing root causes of outages, and developing monitoring systems. Google’s hierarchy of a reliable service are:
- Product: The top of the reliability hierarchy, which indicates that a product is workable and reliable.
- Development: Software engineering and system design work within the company.
- Capacity Planning: Loan balancing ensures that the capacity that was built is being used properly.
- Testing + Release Procedures: After forming an understanding of what went wrong, actively preventing it. Carefully testing products before they are released.
- Postmortem/Root Cause Analysis: Build a culture of blamelessness and addressing a fix to an issue in order to avoid a repeat incident.
- Incident Response: Being on-call, staying in touch with systems, effective troubleshooting, and careful planning before the fact.
- Monitoring: Being aware of problems before the end user notices.
SRE excellence requires experience
The role of a site reliability engineer is best performed by someone with software experience under their belt—it is certainly not a recommended entry-level position. Proper SRE execution requires fluency in software engineering and understanding a system of great scale and complexity.
SRE is a philosophy
A site reliability engineer needs the right mindset for this position. Technical skills are necessary, but a conceptual understanding of operations is key. It is important for SREs to be grounded in traditional software development processes, but there is also a great deal of importance in a holistic understanding of company processes and moving a reliable system forward.
SRE should be a catalyst for change
It should be the job of everyone in the organization to be as reliable as possible, thus applying the important principles of SRE. Apply a reliability model to each team and take the time to discuss how reliability can fit into each team and affect everyone.
Capabilities that scale with your business
Create your next SRE transformation plan with ServiceNow.