English | 中文
CudgX is an AIOps intelligent operation & maintenance engine for the cloud-native era launched by Galaxy Future, it indexes and digitally measures various services through multi-dimensional big-data collection and machine learning training analysis. Based on the deployed training model, the real-time measurement of service quality could achieve the automated and intelligent real-time dynamic scheduling of computing power, storage, network, and other basic resources.
Main Features:
- Support service logging record and aggregate computing;
- Support automatic pressure measurement
- Support real-time capacity evaluation
- Support automatic, intelligent computing-force elastic schedule
- Provide agency access to mesh agent SDK
- Support automatic measurement and Auto Scaling of Web service metric
- With the open platform of metrics definition, partners could configure customized metrics based on the platform.
Weibo | Zhihu | Bilibili | WeChat Official Account | WeCom Communication Group
- System Architecture Diagram
-
System Module Description
-
cudgx-api:
-
predict-rules: Responsible for maintaining Scaling rules.
-
redundancy-keeper: Responsible for maintaining the redundancy of service clusters.
-
-
cudgx-gateway: Responsible for collecting dotting data from
metrics-go
and distributing it to Kafka. -
cudgx-consumer: Responsible for consuming Kafka data, and storing it in clickhouse.
-
metrics-go: CudgX dots SDK.
-
- View Service Monitoring Legends: Redundancy Trends, QPS, Machine Amount
- View record of service scaling
- Example:
-
Configuration Requirements To ensure the stable running of the system, it is recommended that the system has 2 cores and 4 GB memory storage; CudgX has been installed and tested in Linux and MacOS.
-
Environment Dependency Before installing CudgX, please install:
- BridgX: Please install BridgX according to BridgX installation Guide. The intranet deployment environment must be able to communicate with the cloud vendor’s VPC
- SchedulX: Please install SchedulX according to the SchedulX installation Guide. The intranet deployment environment must be able to communicate with the cloud vendor’s VPC
- ComandX: If front-end operations are required, please install ComandX according to the [ComandX installation Guide] (https://github.com/galaxy-future/comandx/blob/main/README.md).
- Installment Steps
-
(1) Download the source code:
- Back-end engineering:
git clone https://github.com/galaxy-future/cudgx.git
-
(2)MacOS(intel chip only) system installation and deployment
-
Back-end deployment, running in CudgX directory
make docker-run-mac
-
-
(3)Linux system installation and deployment
-
1)For users:
-
Back-end deployment, running in CudgX directory
make docker-run-linux
-
-
2)For developers
- Since the project will download the required basic images, it is recommended to place the downloaded source code in a directory that is larger than 10 GB
- Back-end deployment
- CudgX relies on mysql & kafka & clickhouse artifact,
-
If use the built-in mysql & etcd & clickhouse, access the CudgX root directory, and use the following command:
docker-compose up -d //start up CudgX
docker-compose down //stop CudgX -
If already have external mysql & etcd & clickhouse service, go to
cd conf
and change the IP and port configuration information for the corresponding artifacts ofapi.json
gateway.json
consumer.json
, then go to the root directory of CudgX, use the following command:docker-compose up -d api //start up api
docker-compose up -d gateway //tart up gateway
docker-compose up -d consumer //tart up consumer
docker-compose down //stop CudgX
-
- CudgX relies on mysql & kafka & clickhouse artifact,
-
-
(4)Front-end page
- If front-end operations are required, please install ComandX
- After the system runs, enter 'http://127.0.0.1' to view the management console page. The initial username is “root” and the password is “123456”.
- If front-end operations are required, please install ComandX
-
(5)metrics-go
-
1)To complete uploading the data, the target application should dot based on CudgX-SDK metrics-go, we have provided a sample application
cudgx-sample-pi
for testing: Thecudgx-sample-pi
application has been buried based on metrics-go, and the docker imagegalaxyfuture/cudgx-sample-pi
has been pushed to docker hub. -
2)Completing service deployment through SchedulX, see the process of creating SchedulX services.
-
Key Deployment:
-
ComandX Page - Service Deployment - Create Scaling Process – Image Deployment – Services Startup Command:
docker run -d -e CUDGX_SERVICE_NAME=test.cudgx.gf -e CUDGX_CLUSTER_NAME=default -e CUDGX_GATEWAY_URL=http://127.0.0.1:8080 -p 80:8090
Environment Variables Parameter Description:
- CUDGX_SERVICE_NAME: Sverice Name
- CUDGX_CLUSTER_NAME: Cluster Name
- CUDGX_GATEWAY_URL: CudgX-gateway Service IP address
-
ComandX Page-Service Deployment - Create Scaling Process – Traffic Access – Configure SLB ID: Connect to Ali cloud SLB.
-
-
-
3)Rules of Scaling Configuration
- Key Configuration
- ComandX Page – Service Deployment – Scaling Rules
- Key Configuration
-
4)Two Methods are recommended for pressuring {SLB IP}:{SLB PORT}/pi Interface
-
a:During the test, we provided
cudgx-sample-benchmark
application to simulate access traffic. Docker imagesgalaxyfuture/cudgx-sample-benchmark
has been pushed to docker hubdocker run -d --name cudgx_sample_benchmark --network host galaxyfuture/cudgx-sample-benchmark --gf.cudgx.sample.benchmark.sever-address={SLB IP}:{SLB PORT}/pi
-
b:By using opensource interface testing tools, apply pressure to {SLB IP}:{SLB PORT}/pi
-
-
5)ComandX Page- Service List – Cluster Monitoring
- View service monitoring legends: Redundancy Trends, QPS, Machine Amount
- View service scaling records
-
CudgX uses Elastic License 2.0 Agreement for Authorization.
If you want more information about the service, scan the following QR code to contact us: