SRE is a job rule and mindset and a set of engineering best practices to make web products reliable. SREs operate at the intersection between software development and systems engineering to solve operational problems to design, build and run large-scale distributed systems scalably, reliably and efficiently.
SRE core functions include
- Monitoring and metrics. Measuring how the service is actually behaving.
- Incident management response. Noticing and responding effectively to service failures in order to preserve the SLA
- Capacity planning, scaling distributed systems Project future demand to serve the increase of volume requests.
- Continue Integration and Continue Deployment solutions Provide CI/CD solutions to automate the continuous development of new application versions and push it to production systems effortlessly without manual interventions.
- Performance Design, development and engineering related to scalability, isolation, latency, throughput and efficiency.
SREs support the lifecycle of the systems before services go live, SRE help to design, consulting and making recommendation. Once the service is live, SREs maintain them by measuring performance metrics like availability, latency, response times and error rate. Once the service reach end of life, SREs decommission them in a predictable fashion with clear messaging and documentation
The Importance of Documentation
At the beginning of SRE team's existence, the org tends to grows and it heavily relies on the performance of highly skilled individuals on the team. The team develops important operational processes and principles as team's knowledge that are passed verbally to new team members. If these concepts are not written in and documented, information gets lost and procedures can get fragmented and pieces missing. If this is allowed to continue, the understanding of why these principals are in place gets lost and team loses its capacity to stay performant.