Migrating from a Monolith to Microservices (Cloud Next '19)
Вставка
- Опубліковано 30 чер 2024
- Microservices enable development teams to be more agile, but it’s hard to break up an existing monolithic application or to run many services. Come to see a time-lapse refactoring of a fictional monolithic web application into decoupled services. An incremental approach and GCP serverless products for hosting, communication, and storage come to the rescue.
We will explore how to get started and use best practices for your journey to a low-ops, high-velocity future.
Cloud Run → bit.ly/2TSCmFz
Cloud Functions → bit.ly/2K63rFD
Watch more:
Next '19 Serverless Sessions here → bit.ly/Next19Serverless
Next ‘19 All Sessions playlist → bit.ly/Next19AllSessions
Subscribe to the GCP Channel → bit.ly/GCloudPlatform
Speaker(s): Martin Omander, Adam Ross
Session ID: SVR202
product:Cloud Run,Cloud Functions,Cloud - Databases - Firestore,Sheets; fullname:Martin Omander,Adam Ross; event: Google Cloud Next 2019; re_ty: Publish; product: Cloud - General; fullname: Martin Omander, Adam Ross; - Наука та технологія
24:00 What they don't mention here, and what they don't mention almost anywhere, is a really important requirement of the Knative container contract that Cloud Run inherits: containers have dramatically throttled CPU outside the scope of a single request. That means that between starting up your server and handling the first request, you can't do anything. You have to initialize all your variables and everything AFTER you start handling the first request. This is almost impossible to debug if you don't know about it, so it's very frustrating that it's almost never mentioned.
great!
An application also includes user interface and database. More often than not a database already contains few million data. How do these two elements participate with this type of architecture transition
cool!