Xrpc is a framework for creating production quality API services on top of Netty. The framework helps to encapsulate our best practices and provides sane (and safe) defaults.
It currently supports the http/1.1 and the http/2 protocols. It does so interchangeably, i.e your implementation does not need to change and it will automatically respond to a http/1.1 and a http/2 client the same way. The user is free to determine whatever payload they would like, but our recommendation is JSON where you don't control both ends and protobuf (version 3) where you do.
# Building the jar
```shell
$ ./gradlew shadowJar
$ java -jar app.jar
$ ./bin/startPeopleTestServer.sh
$ curl -k -d "bob" -X POST https://localhost:8080/people
--! This demo requires curl with http/2 !-- (see https://simonecarletti.com/blog/2016/01/http2-curl-macosx/)
$ curl -k https://localhost:8080/people
[{"name":"bob"}]
$ curl -k https://localhost:8080/people --http1.1
[{"name":"bob"}]
Run the dino app server to demo proto buffer handling.
$ ./bin/startDinoTestServer.sh
--! This demo requires curl with http/2 !-- (see https://simonecarletti.com/blog/2016/01/http2-curl-macosx/)
$ java -cp app.jar com.nordstrom.xrpc.demo.DinoSetEncoder trex blue | curl -k https://localhost:8080/DinoService/SetDino --data-binary @- -vv
$ java -cp app.jar com.nordstrom.xrpc.demo.DinoGetRequestEncoder trex | curl -k -s https://localhost:8080/DinoService/GetDino --data-binary @-
trexblue
xrpc comes with some built in admin routes. See also Router.java and AdminHandlers.java
/metrics
-> Returns the metrics reporters in JSON format/health
-> Expose a summary of downstream health checks/ping
-> Responds with a 200-OK status code and the text 'PONG'/ready
-> Exposes a Kubernetes or ELB specific healthcheck for liveliness/restart
-> Restart service (should be restricted to approved devs / tooling)/killkillkill
-> Shutdown service (should be restricted to approved devs / tooling)
Please see the contributing guide for details on contributing to this repository.