TSDB-EXPRESS

master

Introduction

A time series database object-relational mapping(ORM) library, support JPA-compatible entity definition and method declaration. After the measurement models definition and corresponding generic type interface declaration, you can query and save influxDB data as a normal java bean.

Requirement

• JDK version 1.7 or later
• maven version 3.0+
• Spring Framework 4.0.0 or later

Features

• It aims to response queries in real-time, which means every data point is indexed as it comes in and is immediately available in queries that should return in < 100ms.
• JPA-compatible query syntax, support operator AND,OR,NOT,BEFORE,AFTER,GREATERTHAN, LESSTHAN etc.
• No redundant XML configuration, annotation type programming.
• Support data source connection pool, improve data access performance.
• Do not need to implement repository operator, just declare corresponding interface and use it.
• Minimum dependency installation.

Prerequisites

• Install InfluxDB version 1.5.2 and chronograf as InfluxDB web UI.

  You can easily setup and run influxDB and chronograf as a docker container service, the deployment and configuration files can be found here.

• Create InfluxDB databases with retention policy.

>influx -precision rfc3339
>CREATE DATABASE "metrics_db3" WITH DURATION 3h REPLICATION 1 NAME "rp_3h"

• Add resource file tsdb.properties in your project.

#tsdb.properties
#optional param, default influxDB
tsdb.datasource.type=influxDB
#optional param, data source connection pool size,  default 1
tsdb.datasource.maxSize=2

#influxDB url
influxdb.address=http://10.88.2.103:8086
influxdb.user=root
influxdb.password=root
influxdb.dbName=metrics_db3
influxdb.rpName=rp_3h
influxdb.maxBatchSize=10
influxdb.maxBatchInterval=100

#optional param, add http transport compression, default false
influxdb.enableGzip=false

Compile && Install

mvn clean install

maven import

<dependency>
        <groupId>dcits</groupId>
        <artifactId>tsdb-express</artifactId>
        <version>1.0-SNAPSHOT</version>
</dependency>

Usage

influxDB measurement model definition

//Cpu.java
@Measurement(name = "cpu")
public class Cpu {

	@Column(name = "time")
	private String time;

	@Tag(name = "host")
	private String host="aaa";


	@Column(name = "idle")
	private Integer idle;

	@Column(name = "user")
	private Integer user;

	@Column(name = "system")
	private Integer system;


	@Override
	public String toString() {
		return "Cpu{" +
				"time='" + time + '\'' +
				", host='" + host + '\'' +
				", idle=" + idle +
				", user=" + user +
				", system=" + system +
				'}';
	}

...
...

or use abstract class

//Memory.java
@Measurement(name = "memory")
public class Memory extends PercentMeasuerment {
}

influxDB repository interface declaration

//Cpu measurement repository
package com.dcits.repo.othermetrics;
public interface RepoCpu extends CustomRepo<Cpu> {

}

Or more complex JPA-compatible interface declaration

package com.dcits.repo.memory;
public interface RepoMemory extends CustomRepo<Memory> {

	public List<Memory> findByIpAddrOrderByTimeDescLimit(String ip, int limit);
	public List<Memory> findByIpAddrLimit(String ip, int limit);
	public List<Memory> findByIpAddrAndTimeBeforeLimit(String ip, String time,int limit);
	public List<Memory> findByIpAddrAndTimeBeforeOrderByTimeDescLimit(String ip, String time,int limit);
	public List<Memory> aggregateByPercentMeanIpAddrAndTimeBeforeGroupByTimeOrderByTimeDescLimit(String ip, String time,String interval, int limit);
}

To make interface above be found and implemented, we add package that contain interfaces within annotation:

package com.dcits.app;


@EnableRepoInterfaceScan({"com.dcits.repo"});
public class Application {
	@Resource
	RepoMemory memExpress;

	...
	...
}

To make this annotation acitve, add this package in your xml:

<context:component-scan base-package="com.dcits.app"></context:component-scan>

JPA method usage examples

    //corresponding influxdb sql: select * from memory where "ipAddr"="192.168.1.100" order by time desc limit 5
    List<Memory> memList = memExpress.findByIpAddrOrderByTimeDescLimit("192.168.1.100",5);
	
    //corresponding influxdb sql: select * from memory where "ipAddr"="192.168.1.100" and time<'2018-06-28T05:34:04.920Z'  order by time desc limit 5
    List<Memory> memList = memExpress.findByIpAddrAndTimeBeforeOrderByTimeDescLimit("192.168.1.100", String.valueOf(before2Min), 5);
	
    //corresponding influxdb sql: select mean(percent) as percent from memory where "ipAddr"="192.168.1.100" and time<'2018-06-28T05:34:04.920Z' group by time(5m) order by time desc limit 5
    List<Memory> memList = memExpress.aggregateByPercentMeanIpAddrAndTimeBeforeGroupByTimeOrderByTimeDescLimit("192.168.1.100", String.valueOf(before2Min),"5m", 5);

Development Instructions

MUST follow the instructions of NOTE section.

• Please install JDK 1.7 before build the project.
• MUST NOT add any domain logics to this project.
• MUST NOT push any jar files, use maven dependency instead.
• MUST NOT push any unnecessary binary files.
• MUST push source code with meaningful message git commit -m "meaningful message".
• MUST import codequality/codestyle-formatter.xml, and format source code (CTRL+SHIFT+F) and organize imports (CTRL+SHIFT+O) before commit.
• MUST use standard JavaDoc Tags on all java source code.
• SHOULD use English in JavaDoc, comments and any source code related resources as possible.
• SHOULD follow Java Coding Conventions and Java Style Guide if you haven't to improve code quality.