💡 DevOps Foundations: Microservices

🪕 1. Microservices In Production

• Decoupling

  • multiple tech stacks
  • services are deployed and scaled independently
  • implementation and changes are isolated
• Increased cohesion (凝聚): different boundaries

• Increased development velocity (速度) and diversity (多元化)

  • smaller services, smaller teams
  • Multiple tech stacks
• Increased operational complexity
• Distributed systems
• Eventual consistency

📑 The 3 common characteristic of well-defined microservices

• Domain-Driven Design
• Loose Coupling, High Cohesion (松耦合，高内聚力)
• Continuous Delivery

🪕 2. Service Readiness

📑 Standardization

• exemplars 范例
• Service Templates: Dropwizard, Governator, Cookiecutter

📑 Unit and integration testing

• Gateway Integration Tests
• Persistence Integration Tests
• Component: In-Process, Out-of-Process Component Tests

📑 End-to-end and contract testing

📑 Contract test

Contract-driven contracts define the expectations of a consumer on a service.

📑 Performance testing: sinon

📑 Continuous integration: git commit

📑 Continuous delivery: Travis CI, Jenkens

In order for continous delivery to be archieved, a build pipeline must create an artifact from source control

📑 Platform-specific and OS artifacts

• What is the recommended approach for structuring code in a microservices architecture?

  Have 1 repository, 1 build pipeline, and 1 artifact per microservice.

• In a microservices architecture, system as a whole should be the focus of performance testing.

📑 Virtual machine and container artifacts: Docker

📑 Versioning: semantic version(^~): major.minor.patch

📑 Basic service discovery

• service registy + basic discovery

• service discovery with DNS

  example of domain-based environments (service_name-environment-organization):

  research-performance.bestitconsulting.com

📑 Dynamic service registries

• 2 Discovery Patterns:

  • Client-side discovery, e.g: Netflix Open-Source Tools
  • Eureka - service registry
  • Ribbon - client-side load balancer
  • Server-side discovery
  • Console and Nginx
  • Kubernetes

• 2 Registration Patterns:

  • Self-registration: heatbeat request
  • Third-party registration

📑 Documentation

• JIRA, sprint story
• Swagger UI: yaml, json, no disquz comments
• gitbook, markdown

📑 Ownership and Conway’s law

🪕 3. Service Resilience, Reliability, and Scalability

📑 Architectural safety measures

📑 Integration methodologies (方法论)

• shared Database
• Synchronous communicatiton
• Asynchronous communicatiton

• Orchestration vs. Choreography 编排与编舞

  • A service collaboration pattern that relies on a central brain to guide and drive business processes
  • describes a system where each part is told what its job is and lets the part itself work out the details.

📑 Synchronous integration technologies

• RPC
• REST: Resources, HATEOAS
• GraphQL: API standard that combines RPC with the REST concept of resources in order to make API interactions more efficient

📑 Asynchronous integration technologies

• Emit events
• Message Broker: Smart endpoints, dumb pipes

📑 Logging

(A) Log Aggregation Tools: (聚合)

• ELK Stack: Elasticsearch, Logstash, and Kibana, Beats.

  “ELK” is the acronym for three open source projects: Elasticsearch, Logstash, and Kibana. Elasticsearch is a search and analytics engine. Logstash is a server‑side data processing pipeline that ingests data from multiple sources simultaneously, transforms it, and then sends it to a “stash” like Elasticsearch. Kibana lets users visualize data with charts and graphs in Elasticsearch.

  The Elastic Stack is the next evolution of the ELK Stack.

• Grafana

(B) Standarize log format

(C) Correlation ID (相关性)

A global unique identifier generated and propagated across service calls for a request.

(D) Distributed Tracing Tools

• Zipkin
• Jaeger

📑 Monitoring

(A) Dashboard: metrics

(B) Synthetic (合成的) Monitoring

(C) Metrics Aggregation Tools: Grafana, Graphite, Prometheus

📑 Alerting

• On-Call Rotation

📑 Incidents (事件)

• Assessment
• Coordination
• Mitigation 减轻
• Resolution
• Follow-Up

📑 Services-level Objectives (SLOs) and error budgets

• SLI: initial indicator.
• A specified target level expressing the desired reliability of a service
• 服务级别目标是指服务提供者向客户作出的服務保證的量化指標。例如軟件提供商向客戶保證一年的時間內有 99.95%的時間應用程序不會出現故障，或是一個月以內 75％的撥打的呼叫中心求助电话将在一分钟内得到答复。這就是一種典型的服务级别目标說明。

• SLOs help determine what engineering work to prioritize:

  • Availability
  • Latency
  • Throughtput
  • Correctness

📑 Capacity planing 容量

• The process to determine the hardware needs of a service.
• Aualitative, Quantitative Growth Scale
• Utilize Autoscaling

🪕 4. Microservices by Example

📑 Overview of KinetEco case study

📑 Greenfield services

• created a service template
• used existing CI infrastructure
• created a shared library for CI/CD functionality and shared base images
• Kubernetes

📑 Splitting the monolith

• seams: 接缝

📑 User-facing authentication and aurthorization

📑 Service-to-service authentication antipatterns

• Authentication and Authorization 认证与授权
• Network-based Protection
• Basic Authentication

📑 Service-to-service authentication

• Single Sign-On
• Client Certificates
• HMAC HTTP Request
• API Keys

📑 Challenges adopting microservices

• API Versioning
• Kubernetes
• Reporting