[7.x] Extrapolate Jaeger transaction count from reported sampling rate (#3722)

[axw]

Backports the following commits to 7.x:

• Extrapolate Jaeger transaction count from reported sampling rate (Extrapolate Jaeger transaction count from reported sampling rate #3722)

[apmmachine]

💚 Build Succeeded

Expand to view the summary

Build stats

• Build Cause: [Branch indexing]

• Start Time: 2020-07-01T18:51:55.135+0000

• Duration: 27 min 29 sec

Test stats 🧪

Test	Results
Failed	0
Passed	3004
Skipped	144
Total	3148

Steps errors

Expand to view the steps failures

• Name: Compress

  • Description: tar --exclude=coverage-files.tgz -czf coverage-files.tgz coverage

  • Duration: 0 min 0 sec

  • Start Time: 2020-07-01T19:08:27.298+0000

  • log

• Name: Compress

  • Description: tar --exclude=system-tests-linux-files.tgz -czf system-tests-linux-files.tgz system-tests

  • Duration: 0 min 0 sec

  • Start Time: 2020-07-01T19:19:21.132+0000

  • log

• Name: Test Sync

  • Description: ./script/jenkins/sync.sh

  • Duration: 3 min 44 sec

  • Start Time: 2020-07-01T19:03:04.385+0000

  • log

[axw]

Verified with BC2. Tested with this program:

package main                                                                            
                                                                                        
import (                                                                                
        "fmt"                                                                           
        "time"                                                                          
                                                                                        
        "github.com/uber/jaeger-client-go"                                              
        "github.com/uber/jaeger-client-go/config"                                       
)                                                                                       
                                                                                        
func main() {                                                                           
        cfg, _ := config.FromEnv()                                                      
        cfg.ServiceName = "schnozz"                                                     
        cfg.Sampler.Type = "remote"                                                     
        cfg.Sampler.Param = 0.1                                                         
        cfg.Sampler.SamplingRefreshInterval = 250 * time.Millisecond                    
                                                                                        
        tracer, _, _ := cfg.NewTracer()                                                 
        for i := 0; i < 1000; i++ {                                                     
                span := tracer.StartSpan("frozzlet")                                    
                fmt.Println("sampled:", span.Context().(jaeger.SpanContext).IsSampled())
                span.Finish()                                                           
                time.Sleep(100 * time.Millisecond)                                      
        }                                                                               
}                                                                                       

Then, confirmed that the number of recorded values per bucket is 10 (1/0.1) times the number of recorded transaction docs:

GET /apm-*/_search
{
  "query": {
    "bool": {
      "filter": [
        {
          "term": {
            "transaction.name": "frozzlet"
          }
        }
      ]
    }
  },
  "size": 0,
  "aggs": {
    "histo_raw": {
      "histogram": {
        "field": "transaction.duration.us",
        "interval": 100
      }
    },
    "histo_histo": {
      "histogram": {
        "field": "transaction.duration.histogram",
        "interval": 100
      }
    }
  }
}

{
  "took" : 3,
  "timed_out" : false,
  "_shards" : {
    "total" : 13,
    "successful" : 13,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 113,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "aggregations" : {
    "histo_histo" : {
      "buckets" : [
        {
          "key" : 0.0,
          "doc_count" : 970
        },
        {
          "key" : 100.0,
          "doc_count" : 30
        },
        {
          "key" : 200.0,
          "doc_count" : 0
        },
        {
          "key" : 300.0,
          "doc_count" : 10
        },
        {
          "key" : 400.0,
          "doc_count" : 0
        },
        {
          "key" : 500.0,
          "doc_count" : 0
        },
        {
          "key" : 600.0,
          "doc_count" : 10
        }
      ]
    },
    "histo_raw" : {
      "buckets" : [
        {
          "key" : 0.0,
          "doc_count" : 97
        },
        {
          "key" : 100.0,
          "doc_count" : 3
        },
        {
          "key" : 200.0,
          "doc_count" : 0
        },
        {
          "key" : 300.0,
          "doc_count" : 1
        },
        {
          "key" : 400.0,
          "doc_count" : 0
        },
        {
          "key" : 500.0,
          "doc_count" : 0
        },
        {
          "key" : 600.0,
          "doc_count" : 1
        }
      ]
    }
  }
}