Главная
Study mode:
on
1
Intro
2
Finagle: Status
3
Finagle: @Twitter
4
Your Server as a Function
5
Configuring Servers and Clients
6
What does the server do?
7
Server: Concurrency Limit
8
Server and Client: Request Timeout
9
What does the client do?
10
Client: Response Classification
11
Client: Retries
12
Client: Retry Budget
13
Client: Retry Backoff
14
Client: Timeouts
15
EWMA (Exp Weighted Moving Avg)
16
Client: Load Balancing via Aperture
17
Client: Circuit Breaking
18
Client: Failure Accrual
19
What else?
Description:
Dive deep into the internals of Finagle, an extensible RPC system for JVM written in Scala, in this 46-minute talk from Scala Days New York 2016. Explore the complex logic behind Finagle's programming model, where servers and clients are viewed as easily composable functions. Discover the inner workings of connection pooling, load balancing, and circuit breaking. Follow along as the speaker walks through each module, explaining what happens when sending a request to a Finagle client and receiving a request by a Finagle server. Learn about server concurrency limits, request timeouts, client response classification, retries, retry budgets, backoff strategies, and load balancing techniques. Gain insights into circuit breaking, failure accrual, and other advanced features that make Finagle a powerful tool for building distributed systems.

Finagle Under the Hood

Scala Days Conferences
Add to list