SRE Mindset
How does a system work? How does it fail?
For SRE, the primary question is not, "How is it supposed to work?" but rather, "How does it really work? How does it really work in production?"
SREs are relentlessly collaborative.
SREs are always searching for good data that can help them improve their reliability efforts.
The mindset should change from "monitor all the things" to "measure reliability from the customer perspective, not the component perspective," from the default of "keep everything up 24–7–365" to "the appropriate level of reliability".
The basic idea is that SRE seeks to improve system reliability collaboratively using objective data.
The first piece of advice is to find your connection to reliability and figure out how to swim in that direction.
SRE's Relationship to Failure and Errors Learning from failure is a core component of the SRE mindset. SRE treats errors as signals.
The SRE mindset differs from most other mindsets in treating failure as an opportunity to learn.
Learning from failure is a core component of the SRE mindset. With an SRE mindset, errors can serve the larger purpose of helping us understand that system.
A distinguishing facet of the SRE mindset is the strong sense of ownership SREs have over the services they run.
SREs can be prone to yak shaving.
How does a system work?
- How will the system work if we scale it?
- How can the system work with less operational load?
- How can the system work reliably for more people?
- How does the system fail when the service/product succeeds?