采样模式

更新时间: 2019-06-21

目前 SOFATracer 提供了两种采样模式,一种是基于 BitSet 实现的基于固定采样率的采样模式;另外一种是提供给用户自定义实现采样的采样模式。下面通过案例来演示如何使用。

本示例基于 tracer-sample-with-springmvc 工程;除 application.properties 之外,其他均相同。

基于固定采样率的采样模式

在 application.properties 中增加采样相关配置项

  1. #采样率 0~100
  2. com.alipay.sofa.tracer.samplerPercentage=100
  3. #采样模式类型名称
  4. com.alipay.sofa.tracer.samplerName=PercentageBasedSampler

验证方式

  • 当采样率设置为100时,每次都会打印摘要日志。
  • 当采样率设置为0时,不打印
  • 当采样率设置为0~100之间时,按概率打印以请求 10 次来验证下结果。

1、当采样率设置为100时,每次都会打印摘要日志

启动工程,浏览器中输入:http://localhost:8080/springmvc ;并且刷新地址10次,查看日志如下:

  1. {"time":"2018-11-09 11:54:47.643","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173568757510019269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":68,"current.thread.name":"http-nio-8080-exec-1","baggage":""}
  2. {"time":"2018-11-09 11:54:50.980","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569097710029269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":3,"current.thread.name":"http-nio-8080-exec-2","baggage":""}
  3. {"time":"2018-11-09 11:54:51.542","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569153910049269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":3,"current.thread.name":"http-nio-8080-exec-4","baggage":""}
  4. {"time":"2018-11-09 11:54:52.061","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569205910069269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-6","baggage":""}
  5. {"time":"2018-11-09 11:54:52.560","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569255810089269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-8","baggage":""}
  6. {"time":"2018-11-09 11:54:52.977","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569297610109269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":1,"current.thread.name":"http-nio-8080-exec-10","baggage":""}
  7. {"time":"2018-11-09 11:54:53.389","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569338710129269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-2","baggage":""}
  8. {"time":"2018-11-09 11:54:53.742","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569374110149269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":1,"current.thread.name":"http-nio-8080-exec-4","baggage":""}
  9. {"time":"2018-11-09 11:54:54.142","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569414010169269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-6","baggage":""}
  10. {"time":"2018-11-09 11:54:54.565","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173569456310189269","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-8","baggage":""}

2、当采样率设置为0时,不打印

启动工程,浏览器中输入:http://localhost:8080/springmvc ;并且刷新地址10次,查看 ./logs/tracerlog/ 目录,没有 spring-mvc-digest.log 日志文件

3、当采样率设置为0~100之间时,按概率打印

这里设置成 20

  • 刷新10次请求
  1. {"time":"2018-11-09 12:14:29.466","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173686946410159846","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-5","baggage":""}
  2. {"time":"2018-11-09 12:15:21.776","local.app":"SOFATracerSpringMVC","traceId":"0a0fe8ec154173692177410319846","spanId":"0.1","request.url":"http://localhost:8080/springmvc","method":"GET","result.code":"200","req.size.bytes":-1,"resp.size.bytes":0,"time.cost.milliseconds":2,"current.thread.name":"http-nio-8080-exec-2","baggage":""}

按 20% 进行采样,测试结果仅供参考

自定义采样模式

SOFATracer 中提供了一个采样率计算的接口 Sampler ,如果你需要定制化自己的采样策略,可通过实现此接口来进行扩展。下面通过一个简单的Demo 来演示如何使用自定义采样模式。

自定义采样规则类

  1. public class CustomOpenRulesSamplerRuler implements Sampler {
  2. private static final String TYPE = "CustomOpenRulesSamplerRuler";
  3. @Override
  4. public SamplingStatus sample(SofaTracerSpan sofaTracerSpan) {
  5. SamplingStatus samplingStatus = new SamplingStatus();
  6. Map<String, Object> tags = new HashMap<String, Object>();
  7. tags.put(SofaTracerConstant.SAMPLER_TYPE_TAG_KEY, TYPE);
  8. tags = Collections.unmodifiableMap(tags);
  9. samplingStatus.setTags(tags);
  10. if (sofaTracerSpan.isServer()) {
  11. samplingStatus.setSampled(false);
  12. } else {
  13. samplingStatus.setSampled(true);
  14. }
  15. return samplingStatus;
  16. }
  17. @Override
  18. public String getType() {
  19. return TYPE;
  20. }
  21. @Override
  22. public void close() {
  23. // do nothing
  24. }
  25. }

在 sample 方法中,用户可以根据当前 sofaTracerSpan 提供的信息来决定是否进行打印。这个案例是通过判断 isServer 来决定是否采样,isServer=true,不采样,否则采样。

在 application.properties 中增加采样相关配置项

  1. com.alipay.sofa.tracer.samplerName.samplerCustomRuleClassName=com.alipay.sofa.tracer.examples.springmvc.sampler.CustomOpenRulesSamplerRuler

相关实验结果,可以自行验证下。