1 Reliable, Scalable, And Maintainable Applications

3 important characteristics

roughly "continue to work correctly even if things go wrong".
Fault-Tolerant - systems which can cope up with faults.

Fault : one component of system deviates from its behaviour
Failure : whole system stops providing the service.

Motive : to prevent faults from becoming failures.

Ability to cope up with increased load

ex: Response time : time between client sending a request and recieving a response

get all response time and sort it from low to high, median is the mid point of that list.

p50: half of user requests are served before median time and half after median time

In other words considering a fixed time as threshold, p50 means 50% of requests take less than the threshold time for reponse time.

p95 : 95% of requests response times are below the threshold
Similarly p99, p999 (99.9%)

Tail Latency: High percentiles of response times

Operability: make it easy for operations teams to keep the system running smoothly
Simplicity: remove as much complexity as possible
Evolvability: Easier to make changes in the future