MicroServices notes: Difference between revisions
From Federal Burro of Information
Jump to navigationJump to search
(→Change) |
|||
| (6 intermediate revisions by the same user not shown) | |||
| Line 35: | Line 35: | ||
==== Data persistance ==== | ==== Data persistance ==== | ||
CAP theorem - In the event of a network partition the trade off between consistency and availability. | |||
==== Infrastructure ==== | ==== Infrastructure ==== | ||
| Line 42: | Line 45: | ||
==== Stateful ==== | ==== Stateful ==== | ||
==== Hybrid ==== | ==== Hybrid ==== | ||
caching layer is hybrid | |||
consider VCache, a system for managing memcache (server) and spymemcached (client) | |||
https://github.com/Netflix/EVCache | |||
lessons: | |||
separate real time from batch | |||
=== Variance === | === Variance === | ||
=== Change === | === Change === | ||
==== Dealing with incidents ==== | |||
Incident -> Resolution -> Review -> remediation -> Analysis -> Best practice -> Automation -> Adoption | |||
=== Production Readiness Reviews === | |||
* Alerts | |||
* Automated Canary Analysis | |||
* Auto Scaling | |||
* Chaos | |||
* Consistent Naming | |||
* ELB Config | |||
* HealthChecks | |||
* Immutable Machine Images | |||
* Squeeze Testing - uwt is this ? | |||
* Timeout, retries, fall back, circuit breaker | |||
== Reading == | |||
; Mastering Chaos - A Netflix Guide to Microservices | |||
: https://www.youtube.com/watch?v=CZ3wIuvmHeM | |||
Latest revision as of 02:20, 9 July 2019
Mastering Chaos - A Netflix Guide to Microservices - Notes
Evolutionary response
- Separation of concern: modularity and encapsulation
- Scalability: horizontal, and partitioning
- Virtualization and Elasticity - automation, and on demand capacity
Dependancy
Use cases:
Intra-Service Requests
can handle non ideal issues
- network issues
- back off
- handling failure.
- can protect against one service failing?
Hystrix
FIT - Fault Injection Testing
Graceful degradation - testing the removal of things, some less important, some more.
Client libraries
Bare bones , or client libraries?
Data persistance
CAP theorem - In the event of a network partition the trade off between consistency and availability.
Infrastructure
Scale Use Cases
Stateless
Stateful
Hybrid
caching layer is hybrid
consider VCache, a system for managing memcache (server) and spymemcached (client)
https://github.com/Netflix/EVCache
lessons:
separate real time from batch
Variance
Change
Dealing with incidents
Incident -> Resolution -> Review -> remediation -> Analysis -> Best practice -> Automation -> Adoption
Production Readiness Reviews
- Alerts
- Automated Canary Analysis
- Auto Scaling
- Chaos
- Consistent Naming
- ELB Config
- HealthChecks
- Immutable Machine Images
- Squeeze Testing - uwt is this ?
- Timeout, retries, fall back, circuit breaker
Reading
- Mastering Chaos - A Netflix Guide to Microservices
- https://www.youtube.com/watch?v=CZ3wIuvmHeM
