RTAnomaly

Detecting anomalous traces of microservice system.

Dependencies

pip install -r requirements.txt

Dataset

Training set: train_ticket/train/train_dataset.txt

Test normal traces: train_ticket/test_normal/test_dataset.txt

Test anomalous traces: train_ticket/test_abnormal/test_dataset_fault.txt

Usage

./run.sh 

Train Ticket：A Benchmark Microservice System

The project is a train ticket booking system based on microservice architecture which contains 41 microservices. The programming languages and frameworks it used are as below.

• Java - Spring Boot, Spring Cloud
• Node.js - Express
• Python - Django
• Go - Webgo
• DB - Mongo、MySQL

You can get more details at Wiki Pages.

Service Architecture Graph

Quick Start

We provide k8s deployment to quickly deploy our application: Using Kubernetes.

Using Kubernetes

Here is the steps to deploy the Train Ticket onto any existing Kubernetes cluster.

Presequisite

• An existing Kubernetes cluster
• Helm supported, you can see https://helm.sh/docs/helm/helm_install/ for helm install
• PVC supported, you can see https://openebs.io/docs/2.12.x/user-guides/installation for localPV support.

1. Clone the Repository

git clone --depth=1 https://github.com/FudanSELab/train-ticket.git 
cd train-ticket/

2. Deploy the application

For Quick Start

make deploy

Note: if you want specify namespace, set Namespace paramter:

make deploy Namespace=yournamespace

Deploy Mysql Clusters For Each Services

make deploy DeployArgs="--independent-db"

With Moinitorig

make deploy DeployArgs="--with-monitoring"

With Distributed Tracing

make deploy DeployArgs="--with-tracing"

Deploy All

make deploy DeployArgs="--all"

Customise Deployment

You can freely combine parameters for custom deployment， for example, deploy with monitoring and tracing:

make deploy DeployArgs="--with-tracing --with-monitoring"

Reset Deployment

make reset-deploy
# if you specify namespace when deploy, set namespace as well when reset
# make reset-deploy Namespace=yournamespace

3. Run kubectl get pods to see pods are in a ready state

4. Visit the Train Ticket web page at http://[Node-IP]:32677.