Summary of the video DevOps Vs SRE by Google

DevOpsVsSRE.txt

DevOps Vs. SRE

DevOps Manifesto
1. Reduce Organizational Silos. (Cross-functionalities)
2. Accept Failure as Normal. (Be ready for failure and plan a backup)
3. Implement Gradual Change. (Deploy small incremental changes)
4. Leverage Tooling & Automation. (Automate repeatable things)
5. Measure everything. 

Above are abstract ideas.

class SRE implements DevOps

DevOps is like an abstract class or an interface in programming whereas SRE is a concrete implementation of that class; it tells the hows. 
1. Reduce Organizational Silos. (Share ownership by using same shared set of toolings)
2. Accept Failure as Normal. (SLO: Service Level Objectives, commonly determine how much fault can be expected)
3. Implement Gradual Change. (Deploy small incremental changes)
4. Leverage Tooling & Automation. (TOIL: invest in tooling which brings long term value)
5. Measure everything. 

SLI: Service Level Indication (Product, SRE, SWE)
Request Latency
Requests per second
Failures per request

At any given moment is it healthy or not in the system.

SLO: Service Level Objectives (SRE, Product)
Binding target for a collection of SLIs
If we were to integrate all of those points over a time period, that gives the SLO.

SLI measures up or down and the SLO says how much up or down can we have in a particular time period like a quarter or a half. 

SLA: Service Level Agreement (Sales, Customer)
Business agreement between a customer and service provider typically based on SLOs

SLIs drive SLOs which inform SLAs.

Acceptable risk dictates the SLO.

TOIL
Manual, repetitive, automatable, tactical, devoid of long term.