add pprof profiller for debugging (#114)

* feat: add pprof profiller for debugging

* feat: load profiller on a more generic way

* fix: lint issues

* feat: adding docs

* feat: adding some extra lines about performance

Author: emmanuelm41

pkg/zprofiller/README.md

@@ -0,0 +1,111 @@
+# ZProfiller
+
+## Overview
+
+`zprofiller` is a Go package designed to facilitate the integration of Go's built-in `pprof` profiling into web applications using the Chi router. It offers a structured way to expose profiling endpoints to monitor and analyze the performance of Go applications in both development and production environments.
+
+## Features
+
+- **Easy Integration**: Seamlessly integrates with the Chi router.
+- **Automatic Profiling Endpoints**: Automatically registers standard `pprof` endpoints.
+- **Custom Configuration**: Allows customization of server timeouts and logging.
+
+## Getting Started
+
+### Installation
+
+To use `zprofiller` in your project, ensure you have Go installed and your workspace is set up, then add `zprofiller` to your dependencies:
+
+```
+go get -u github.com/zondax/zprofiller
+```
+
+### Integration
+
+Integrate `zprofiller` into your Go application:
+
+1. **Import the Package**
+
+   ```go
+   import (
+       "github.com/zondax/zprofiller"
+   )
+   ```
+
+2. **Create a Config Object**
+
+   ```go
+   config := &zprofiller.Config{
+       ReadTimeOut:  5 * time.Second,
+       WriteTimeOut: 5 * time.Second,
+       Logger:       logger.NewLogger(),
+   }
+   ```
+
+3. **Instantiate zprofiller**
+
+   ```go
+   profiler := zprofiller.New(nil, config)
+   ```
+
+4. **Run the Profiler**
+
+   ```go
+   err := profiler.Run(":9999")
+   if err != nil {
+       log.Fatalf("Failed to start profiler: %v", err)
+   }
+   ```
+
+### Usage
+
+Access the profiling endpoints at `http://localhost:<port>/debug/pprof/`, where `<port>` is the port you specified.
+
+## Viewing pprof Results on the Web
+
+### Accessing pprof via Web Browser
+
+Navigate to:
+
+```
+http://localhost:<port>/debug/pprof/
+```
+
+This index page links to profiles like Heap, Goroutine, Threadcreate, Block, and Mutex.
+
+### Visualizing Profiles
+
+Use tools like `Go Tool Pprof` or `Graphviz` for deeper analysis:
+
+- **Go Tool Pprof**:
+
+  ```
+  go tool pprof -http=:8081 http://localhost:<port>/debug/pprof/profile
+  ```
+
+  This command downloads the CPU profile data from your application and opens it in an interactive web interface on `http://localhost:8081`.
+
+- **Graphviz**:
+
+  ```
+  sudo apt-get install graphviz
+  go tool pprof -http=:8081 --graph http://localhost:<port>/debug/pprof/profile
+  ```
+
+### Online Tools and Extensions
+
+Consider using online tools or browser extensions like **pprof++** for Chrome for in-browser visualization of pprof data.
+
+## Performance Considerations
+
+When integrating profiling tools such as `pprof` into your application, it is essential to consider the potential impact on performance:
+
+- **Resource Usage**: Profiling operations can consume significant CPU and memory resources, particularly when capturing and analyzing high-frequency data such as CPU profiles.
+- **Production Use**: While `pprof` can be invaluable for diagnosing issues in production, it should be enabled selectively. Consider using environment variables or configuration files to control access to profiling endpoints.
+- **Sampling Rate**: Adjust the sampling rate of profiles according to the performance impact and the level of detail required. Lower rates can reduce overhead but may miss critical details.
+- **Security**: Exposing profiling information can introduce security risks. Ensure that profiling endpoints are protected with authentication mechanisms and are only accessible by authorized personnel.
+- **Impact Measurement**: Continuously monitor the impact of enabling profiling on your system’s response times and resource usage. Disable profiling when not needed to avoid unnecessary overhead.
+
+## Security Considerations
+
+Ensure that access to profiling endpoints is secured, especially in production environments, to protect sensitive application data.

pkg/zprofiller/zprofiller.go

@@ -0,0 +1,91 @@
+package zprofiller
+
+import (
+	"fmt"
+	"github.com/go-chi/chi/v5"
+	"github.com/zondax/golem/pkg/logger"
+	"github.com/zondax/golem/pkg/metrics"
+	"net/http"
+	"net/http/pprof"
+	pprofRuntime "runtime/pprof"
+	"time"
+)
+
+const (
+	defaultAddress = ":8888"
+	defaultTimeOut = 240000
+)
+
+type Config struct {
+	ReadTimeOut  time.Duration
+	WriteTimeOut time.Duration
+	Logger       *logger.Logger
+}
+
+type zprofiller struct {
+	router *chi.Mux
+	config *Config
+}
+
+type ZProfiller interface {
+	Run(addr ...string) error
+}
+
+func (c *Config) setDefaultValues() {
+	if c.ReadTimeOut == 0 {
+		c.ReadTimeOut = time.Duration(defaultTimeOut) * time.Millisecond
+	}
+
+	if c.WriteTimeOut == 0 {
+		c.WriteTimeOut = time.Duration(defaultTimeOut) * time.Millisecond
+	}
+
+	if c.Logger == nil {
+		l := logger.NewLogger()
+		c.Logger = l
+	}
+}
+
+func New(_ metrics.TaskMetrics, config *Config) ZProfiller {
+	if config == nil {
+		config = &Config{}
+	}
+
+	config.setDefaultValues()
+
+	router := chi.NewRouter()
+	router.HandleFunc("/debug/pprof/", pprof.Index)
+	router.HandleFunc("/debug/pprof/cmdline", pprof.Cmdline)
+	router.HandleFunc("/debug/pprof/profile", pprof.Profile)
+	router.HandleFunc("/debug/pprof/symbol", pprof.Symbol)
+	router.HandleFunc("/debug/pprof/trace", pprof.Trace)
+
+	for _, profile := range pprofRuntime.Profiles() {
+		router.Handle(fmt.Sprintf("/debug/pprof/%s", profile.Name()), pprof.Handler(profile.Name()))
+	}
+
+	zr := &zprofiller{
+		router: router,
+		config: config,
+	}
+
+	return zr
+}
+
+func (r *zprofiller) Run(addr ...string) error {
+	address := defaultAddress
+	if len(addr) > 0 {
+		address = addr[0]
+	}
+
+	r.config.Logger.Infof("Start profiller server at %v", address)
+
+	server := &http.Server{
+		Addr:         address,
+		Handler:      r.router,
+		ReadTimeout:  r.config.ReadTimeOut,
+		WriteTimeout: r.config.WriteTimeOut,
+	}
+
+	return server.ListenAndServe()
+}