Springboot使用influxDB时序数据库的实现

引入依赖

项目中使用influxdb-java,在pom文件中添加如下依赖(github地址:https://github.com/influxdata/influxdb-java):

				?

									<dependency>

									    <groupId>org.influxdb</groupId>

									    <artifactId>influxdb-java</artifactId>

									    <version>2.15</version>

									</dependency>

application.yaml文件配置如下所示(请按照实际情况填写):

				?

									spring:

									  influx:

									    url: *

									    password: admin

									    user: 123

									    database: log_management

配置

(1) 创建配置类

				?

									@Configuration

									public class InfluxDbConfig {

									    @Value("${spring.influx.url:''}")

									    private String influxDBUrl;

									    @Value("${spring.influx.user:''}")

									    private String userName;

									    @Value("${spring.influx.password:''}")

									    private String password;

									    @Value("${spring.influx.database:''}")

									    private String database;

									    @Bean

									    public InfluxDbUtils influxDbUtils() {

									        return new InfluxDbUtils(userName, password, influxDBUrl, database, "");

									    }

									}

				?

									@Data

									 public class InfluxDbUtils {

									    private String userName;

									    private String password;

									    private String url;

									    public String database;

									    private String retentionPolicy;

									    // InfluxDB实例

									    private InfluxDB influxDB;

									    // 数据保存策略

									    public static String policyNamePix = "logRetentionPolicy_";

									    public InfluxDbUtils(String userName, String password, String url, String database,

									                         String retentionPolicy) {

									        this.userName = userName;

									        this.password = password;

									        this.url = url;

									        this.database = database;

									        this.retentionPolicy = retentionPolicy == null || "".equals(retentionPolicy) ? "autogen" : retentionPolicy;

									        this.influxDB = influxDbBuild();

									    }

									    /**

									     * 连接数据库 ，若不存在则创建

									     *

									     * @return influxDb实例

									     */

									    private InfluxDB influxDbBuild() {

									        if (influxDB == null) {

									            influxDB = InfluxDBFactory.connect(url, userName, password);

									        }

									        try {

									            createDB(database);

									            influxDB.setDatabase(database);

									        } catch (Exception e) {

									            log.error("create influx db failed, error: {}", e.getMessage());

									        } finally {

									            influxDB.setRetentionPolicy(retentionPolicy);

									        }

									        influxDB.setLogLevel(InfluxDB.LogLevel.BASIC);

									        return influxDB;

									    }

									 }

构建实体类

InfluxDB中,measurement对应于传统关系型数据库中的table(database为配置文件中的log_management).
InfluxDB里存储的数据称为时间序列数据,时序数据有零个或多个数据点.
数据点包括time(一个时间戳)，measurement(例如logInfo)，零个或多个tag，其对应于level,module,device_id),至少一个field(即日志内容,msg=something error).
InfluxDB会根据tag数值建立时间序列(因此tag数值不能选取诸如UUID作为特征值,易导致时间序列过多,导致InfluxDB崩溃),并建立相应索引,以便优化诸如查询速度.

				?

									@Builder

									@Data

									@Measurement(name = "logInfo")

									public class LogInfo {

									    // Column中的name为measurement中的列名

									    // 此外,需要注意InfluxDB中时间戳均是以UTC时保存,在保存以及提取过程中需要注意时区转换

									    @Column(name = "time")

									    private String time;

									    // 注解中添加tag = true,表示当前字段内容为tag内容

									    @Column(name = "module", tag = true)

									    private String module;

									    @Column(name = "level", tag = true)

									    private String level;

									    @Column(name = "device_id", tag = true)

									    private String deviceId;

									    @Column(name = "msg")

									    private String msg;

									}

保存数据

以下代码为单条日志保存,influxdb-java亦支持批量保存(因为与InfluxDB通讯均是通过http,因此建议批量保存以减少性能损耗).

				?

									LogInfo logInfo = LogInfo.builder()

									    .level(jsonObject.getString("level"))

									    .module(module)

									    .deviceId(deviceId)

									    .msg(jsonObject.getString("msg"))

									    .build();

									Point point = Point.measurementByPOJO(logInfo.getClass())

									    .addFieldsFromPOJO(logInfo)

									    .time(jsonObject.getLong("time"), TimeUnit.MILLISECONDS)

									    .build();

									// 出于业务考量,设备可以设置不同的保存策略(策略名为固定前缀+设备ID)

									influxDB.write(influxDBUtils.database, InfluxDbUtils.policyNamePix + deviceId, point);

查询数据

因为代码与业务耦合比较厉害,因此此处仅截选做概要示范.

				?

									// InfluxDB支持分页查询,因此可以设置分页查询条件

									String pageQuery = " LIMIT " + request.getPageSize() + " OFFSET " + ((request.getPageNum() - 1) * request.getPageSize());

									// 此处查询所有内容,如果

									String queryCmd = "SELECT * FROM "

									    // 查询指定设备下的日志信息

									    // 要指定从 RetentionPolicyName(保存策略前缀+设备ID).measurement(logInfo) 中查询指定数据)

									    + InfluxDbUtils.policyNamePix + request.getDeviceId() + "." + "logInfo"

									    // 添加查询条件(注意查询条件选择tag值,选择field数值会严重拖慢查询速度)

									    + queryCondition

									    // 查询结果需要按照时间排序

									    + " ORDER BY time DESC"

									    // 添加分页查询条件

									    + pageQuery;