CodeBlanks Logo Icon CodeBlanks
Login Register

Metrics

Learn the Prometheus-style metric concept.

Metrics

Metrics track a system's numeric state over time. There are three basic types: counter (only goes up: total requests), gauge (goes up/down: active connections) and histogram (distribution of values: response times).

istekSayisi.Inc()
aktifBaglanti.Set(12)
sure.Observe(0.043)

Systems such as Prometheus scrape these values from a /metrics endpoint at regular intervals.

Why a histogram?

Average response time is misleading: if 99 requests take 10 ms and one takes 5 seconds, the average looks like 60 ms but one user has a terrible experience. A histogram puts durations into buckets and lets you compute percentiles like p95 and p99. The standard library's expvar offers simple counters; production usually uses the Prometheus client or OpenTelemetry.

Practice expvar counter

Create a counter with expvar.NewInt("istekler"), call Add(1) five times in a loop and print Value().

Starter code

main.go
package main

import (
    "expvar"
    "fmt"
)

func main() {
    // buraya yaz
}

Sample solution

main.go
package main

import (
    "expvar"
    "fmt"
)

func main() {
    istekler := expvar.NewInt("istekler")
    for i := 0; i < 5; i++ {
        istekler.Add(1)
    }
    fmt.Println(istekler.Value())
}

Expected output

5

Questions and answers

Review the lesson questions together with their answers. To test yourself, solve the lesson interactively.

1 Which metric type tracks an in-memory queue length?

  • Gauge Correct answer
  • Counter
  • Histogram
  • Log

Explanation: Queue length can go up and down; a counter only increases. Instantaneous values are measured with a gauge.

2 Increment the counter

Question 
var istekler = expvar.NewInt("istekler")

istekler.____(1)
Answer 
var istekler = expvar.NewInt("istekler")

istekler.Add(1)

Explanation: An expvar.Int counter is increased with Add; the value is published at /debug/vars.

3 Measurement and type

Term Meaning
Total requests Counter
Active connections Gauge
Response-time distribution Histogram
p99 Percentile

All lessons in Advanced

1 Reflection

  1. The reflect package
  2. reflect.Type
  3. reflect.Value
  4. Reflecting struct fields
  5. Dynamic call
  6. Summary - Reflection

2 unsafe

  1. The unsafe package
  2. unsafe.Pointer
  3. Sizeof and Alignof
  4. Offsetof
  5. unsafe and slice
  6. Summary - unsafe

3 cgo

  1. What is cgo?
  2. Calling C code
  3. C data types
  4. Go-C conversion
  5. cgo and memory
  6. Summary - cgo

4 plugin

  1. The plugin package
  2. Building a plugin
  3. Loading a plugin
  4. Plugin symbol
  5. Plugin and error
  6. Summary - plugin

5 database/sql

  1. The database/sql package
  2. sql.Open
  3. Query and Exec
  4. Prepared statement
  5. Transaction
  6. Summary - database/sql

6 Middleware

  1. The middleware concept
  2. HTTP middleware
  3. Middleware chain
  4. Logging middleware
  5. Auth middleware
  6. Summary - Middleware

7 gRPC

  1. What is gRPC?
  2. Protobuf definition
  3. Compiling with protoc
  4. gRPC server
  5. gRPC client
  6. Summary - gRPC

8 Project: Microservice

  1. Service skeleton
  2. Database layer
  3. HTTP API
  4. Docker and deployment

9 Docker

  1. Dockerfile
  2. Multi-stage build
  3. scratch image
  4. docker-compose
  5. Build args and labels
  6. Summary - Docker

10 pprof

  1. What is pprof?
  2. CPU profiling
  3. Memory profiling
  4. go test -cpuprofile
  5. The trace tool
  6. Summary - pprof

11 Observability

  1. log/slog
  2. Metrics
  3. Tracing
  4. OpenTelemetry
  5. Health check
  6. Summary - Observability

12 Advanced final

  1. Challenge 1 - Reflection
  2. Challenge 2 - Microservice
  3. Challenge 3 - Performance
  4. Challenge 4 - Database