A Service Level Expectation (SLE) is an internal commitment to the reliability a system should deliver. Unlike a contractual SLA, it creates no external penalty; it gives the team a standard for recognizing and improving reliability gaps. section 12.3.1
Make the expectation measurable and use it to guide operational improvement:
- Set targets such as p99 API latency under 200 milliseconds, recovery from a single service failure within five minutes, or change failure rate below 5%. section 12.3.1
- Measure against the targets weekly. Treat each breach as a signal to investigate and improve, not as a reason to wait for a customer complaint. section 12.3.1
Table 12.2 Uptime targets and the cost of hesitation
|
SLA target
|
Allowed downtime per year
|
Allowed per month
|
35 min incident consumes
|
|
99.9% (three nines)
|
8 hours 46 min
|
43.8 min
|
80% of monthly budget
|
|
99.95%
|
4 hours 23 min
|
21.9 min
|
160% of monthly budget
|
|
99.99% (four nines)
|
52.6 min
|
4.38 min
|
8 months of annual budget
|
Compare how a 35-minute outage consumes allowed downtime under different SLA targets; the comparison illustrates why teams can set internal reliability expectations even without a contract. section 12.3.1
