commit | author | age
|
f92636
|
1 |
Customer |
JS |
2 |
======== |
|
3 |
|
|
4 |
This is the customer microservice, part of the Istio Tutorial demo. Even though this microservice is meant to be executed within a Container on a Pod on Kubernetes/OpenShift, it can still be executed on bare metal. |
|
5 |
|
|
6 |
This is a regular Wildfly Swarm application, with OpenTracing and Jaeger dependencies to provide distributed tracing capabilities. |
|
7 |
|
|
8 |
Running on the local machine |
|
9 |
============================ |
|
10 |
|
|
11 |
To run this service for development purposes on your own machine, execute: |
|
12 |
|
|
13 |
```bash |
|
14 |
JAEGER_SERVICE_NAME=customer mvn wildfly-swarm:run |
|
15 |
``` |
|
16 |
|
|
17 |
The environment variable `JAEGER_SERVICE_NAME` is required, as the Jaeger Tracer is embedded into our application and expects a service name to be specified. The example should work, however, even when an installation of Jaeger is *not* available. |
|
18 |
|
|
19 |
If you do prefer to have a local Jaeger instance running to see the traces, the easiest is to start via a Docker |
|
20 |
container: |
|
21 |
|
|
22 |
```bash |
|
23 |
docker run \ |
|
24 |
--rm \ |
|
25 |
-p5775:5775/udp \ |
|
26 |
-p6831:6831/udp \ |
|
27 |
-p6832:6832/udp \ |
|
28 |
-p16686:16686 \ |
|
29 |
-p14268:14268 \ |
|
30 |
jaegertracing/all-in-one:1.3 |
|
31 |
``` |
|
32 |
|
|
33 |
The default configuration for the Jaeger tracer samples only a small portion of the requests. To trace every incoming request and report the spans to the log file, export the following environment variables and start the application again: |
|
34 |
|
|
35 |
```bash |
|
36 |
export JAEGER_REPORTER_LOG_SPANS=true |
|
37 |
export JAEGER_SAMPLER_TYPE=const |
|
38 |
export JAEGER_SAMPLER_PARAM=1 |
|
39 |
``` |
|
40 |
|
|
41 |
To test, call http://localhost:8280/ |
|
42 |
|
|
43 |
``` |
|
44 |
$ curl http://localhost:8280/ |
|
45 |
customer => preference => recommendation v1 from 'caju': 3 |
|
46 |
``` |
|
47 |
|
|
48 |
![Trace View](trace.png) |
|
49 |
|
|
50 |
Running on OpenShift |
|
51 |
==================== |
|
52 |
|
|
53 |
The following commands will build a Docker image containing the application, create a Kubernetes `Deployment` and a corresponding `Service`, so that other services can discover the pods via the service name. |
|
54 |
|
|
55 |
```bash |
|
56 |
mvn clean package |
|
57 |
docker build -t example/customer . |
|
58 |
docker images | grep customer |
|
59 |
oc apply -f ../../kubernetes/Deployment.yml |
|
60 |
oc apply -f ../../kubernetes/Service.yml |
|
61 |
oc expose service customer |
|
62 |
``` |
|
63 |
|
|
64 |
The last command will expose the service to the outside world, allowing you to make an HTTP call directly from your host machine: |
|
65 |
|
|
66 |
``` |
|
67 |
curl http://customer-tutorial.127.0.0.1.nip.io/ |
|
68 |
``` |