Reliable / Reliability
-
System to continue to work correctly, even when things fail. When one
component dies, the redundant component can take its place while.
What things can go wrong?
| 1. Hardware Failures |
a. Hard Disk Failures: Add redundancy Disk, RAID configuration b. Power Failures: Use Generators or Dual power supply |
| 2. Software Failures |
a. High CPU, Memory, n/w BW: OOM killer, Autoscaling b. Outdated Software: Rolling updates c. Leaks, Ransomeware attacks How to void software failures? Good design patterns, use tested libraries, Monitoring, logging, alerting. |
Scalable / Scalability
-
Means System behaves normally under increased load
How to make system Scalable?
1. HPA (Horizontal Pod Autoscaling)
2. DB Scaling: Sharding/Partitioning DB
Maintainable / Maintainability
-
Over time, many different people will work on the system (engineering
and operations, both maintaining current behavior and adapting the
system to new use cases), and they should all be able to work on it
productively.