Google Professional Cloud Architect Certification: 7. Designing for Reliability - Important Things To Know

Reliability is a property of systems that measures the probability that a service will be available for a period of time. Several factors contribute to creating and maintaining reliable systems, including monitoring systems using metrics, logs, and alerts.
Continuous integration and continuous deployment are commonly used practices for managing the release of code. These practices reduce risk by emphasizing the frequent release of small changes. The use of automation helps to reduce the risk that a bad deployment will disrupt services for too long.
Systems reliability engineering is a set of practices that incorporates software engineering practices with operations management. These practices recognize that failures will occur, and the best way to deal with those failures is to set well-defined service-level objectives and service-level indicators, monitor systems to detect indications of failure, and learn from failures using techniques such as post-mortem analysis. Systems should be architected to anticipate problems, such as overloading and cascading failures. Testing is an essential part of promoting highly reliable systems.
 

1. Know the role of monitoring, logging, and alerting in maintaining reliable systems. Monitoring collects metrics, which are measurements of key attributes of a system, such as utilization rates. Metrics are often analyzed in a time series. Logging is used to record significant events in an application or infrastructure component. Alerting is the process of sending notifications to human operators when some condition is met indicating that a problem needs human intervention. Conditions are often of the form that a resource measurement exceeds some threshold for a specified period of time.
2. Understand continuous deployment and continuous integration. CD/CI is the practice of releasing code soon after it is completed and after it passes all tests. This allows for rapid deployment of code. Continuous integration is the practice of incorporating code changes into baseline code frequently. Code is kept in a version control repository that is designed to support collaboration among multiple software engineers.
3. Know the different kinds of tests that are used when deploying code. These include unit tests, integration tests, acceptance tests, and load testing. Unit tests check the smallest unit of functional code. Integration tests check that a combination of units function correctly together. Acceptance tests determine whether code meets the requirements of the system. Load testing measures how well the system responds to increasing levels of load.
4. Understand that systems reliability engineering is a practice that combines software engineering practices with operations management to reduce risk and increase the reliability of systems. The core tenets of systems reliability engineering include the following:
Automating systems operations as much as possible
5. Understanding and accepting risk and implementing practices that mitigate risk
6. Learning from incidents
7. Quantifying service-level objectives and service-level indicators
8. Measuring performance
9. Know that systems reliability engineering includes design practices, such as planning for overload, cascading failures, and incident response. Overload is when a workload on a system exceeds the capabilities of the system to process the workload in the time allowed. Ways of dealing with overload include load shedding, degrading service, and upstream or client throttling. Cascading failures occur when a failure leads to an action that causes further failures. An example is that in response to a failure in a server in a cluster, a load balancer shifts additional workload to the remaining healthy servers, causing them to fail due to overload. Incident response is the practice of controlling failures by using a structured process to identify the cause of a problem, correct the problem, and learn from the problem.
10. Know testing is an important part of reliability engineering. There are several kinds of tests, and all should be used to improve reliability. Testing for reliability includes practices used in CI/CD but adds others as well, particularly stress testing. These tests may be applied outside of the CI/CD process.