learning-go-for-sre-interview/01-url-health-checker/main.go

// Exercise 01 — Concurrent URL Health Checker
//
// ASSIGNMENT
// ----------
// Implement CheckAll so that it:
//
//  1. Checks every URL in `urls` with an HTTP GET.
//  2. Runs AT MOST `limit` checks concurrently (bounded concurrency).
//  3. Uses a per-request timeout of 5s AND respects the global `ctx`
//     (if ctx is cancelled, pending/unstarted checks must return ctx.Err()).
//  4. Returns one Result per URL, at the SAME index as the input URL.
//  5. Reuses a single http.Client and properly drains+closes response bodies
//     so connections are reused.
//
// Run the tests:   go test -race ./01-url-health-checker/
// Run the demo:    go run ./01-url-health-checker/
package main

import (
	"context"
	"fmt"
	"io"
	"io/ioutil"
	"net/http"
	"sync"
	"time"
)

type Result struct {
	URL     string
	Status  int
	Latency time.Duration
	Err     error
}

// CheckAll checks all URLs with at most `limit` in flight at once.
// TODO: implement me.
func CheckAll(ctx context.Context, urls []string, limit int) []Result {
	results := make([]Result, len(urls))
	//   - semaphore (buffered channel) or worker pool for the concurrency limit
	sem := make(chan struct{}, limit)
	//   - sync.WaitGroup to wait for completion
	var wg sync.WaitGroup
	//   - http.NewRequestWithContext for cancellation
	client := &http.Client{Timeout: 5 * time.Second}

	for i, u := range urls {
		wg.Add(1)
		go func(i int, u string) {
			defer wg.Done()
			// wait for semaphore or ctx cancellation
			select {
			case sem <- struct{}{}:
				defer func() { <-sem }()
			case <-ctx.Done():
				results[i] = Result{URL: u, Err: ctx.Err()}
				return
			}

			start := time.Now()
			req, err := http.NewRequestWithContext(ctx, http.MethodGet, u, nil)
			if err != nil {
				results[i] = Result{URL: u, Err: err}
				return
			}
			resp, err := client.Do(req)
			if err != nil {
				results[i] = Result{URL: u, Err: err, Latency: time.Since(start)}
				return
			}
			defer resp.Body.Close()
			io.Copy(ioutil.Discard, resp.Body)
			results[i] = Result{URL: u, Err: err, Status: resp.StatusCode, Latency: time.Since(start)}
		}(i, u)
	}

	//   - write results[i] from goroutine i (why is this safe without a mutex?)
	_ = ctx
	_ = limit
	wg.Wait()
	return results
}

func main() {
	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
	defer cancel()
	// Pick ONE of the url sets below (uncomment it, comment the others).
	// Tip: run with a bigger `limit` and compare total wall-clock time.

	// Option A — the original mix: OK, HTTP 500, DNS failure.
	urls := []string{"https://example.com", "https://httpbin.org/status/500", "https://nope.invalid"}

	//region Not-selected-options

	// Option B — all succeed (2xx). Good for measuring how `limit` affects
	// total time: with limit=1 it's the sum of latencies, with limit=8 ~max.
	// urls := []string{
	// 	"https://example.com",
	// 	"https://www.google.com",
	// 	"https://www.cloudflare.com",
	// 	"https://github.com",
	// 	"https://go.dev",
	// 	"https://pkg.go.dev",
	// 	"https://httpbin.org/status/200",
	// 	"https://httpbin.org/get",
	// }

	// Option C — HTTP-level failures (connection succeeds, status is bad).
	// Err should be nil for these; the caller decides what a "healthy" status is.
	// urls := []string{
	// 	"https://httpbin.org/status/200", // control
	// 	"https://httpbin.org/status/301", // redirect — client follows it by default, so you'll see the final status
	// 	"https://httpbin.org/status/401",
	// 	"https://httpbin.org/status/403",
	// 	"https://httpbin.org/status/404",
	// 	"https://httpbin.org/status/429",
	// 	"https://httpbin.org/status/500",
	// 	"https://httpbin.org/status/502",
	// 	"https://httpbin.org/status/503",
	// 	"https://example.com/definitely-not-here", // real site, 404
	// }

	// Option D — network/transport failures (Err != nil, Status == 0).
	// Each fails for a DIFFERENT reason — print r.Err and read the message.
	// urls := []string{
	// 	"https://example.com",                    // control
	// 	"https://nope.invalid",                   // DNS: .invalid TLD never resolves
	// 	"https://this-host-does-not-exist-12345.com", // DNS: NXDOMAIN
	// 	"http://localhost:1",                     // TCP: connection refused (nothing listens on port 1)
	// 	"http://10.255.255.1",                    // TCP: unroutable address → connect timeout (hits the 5s per-request timeout)
	// 	"https://expired.badssl.com",             // TLS: expired certificate
	// 	"https://self-signed.badssl.com",         // TLS: self-signed certificate
	// 	"https://wrong.host.badssl.com",          // TLS: hostname mismatch
	// 	"https://untrusted-root.badssl.com",      // TLS: unknown CA
	// 	"ftp://example.com",                      // unsupported protocol scheme
	// 	"not a url at all",                       // request creation fails (invalid URL)
	// }

	// Option E — slow responses: exercises the 5s per-request timeout AND the
	// global ctx. /delay/N sleeps N seconds server-side before replying.
	// With limit=2 and a 10s ctx, watch which ones finish, which time out
	// per-request, and which never start and return ctx.Err().
	// urls := []string{
	// 	"https://httpbin.org/delay/1",  // OK
	// 	"https://httpbin.org/delay/3",  // OK
	// 	"https://httpbin.org/delay/8",  // exceeds 5s per-request timeout
	// 	"https://httpbin.org/delay/8",  // exceeds 5s per-request timeout
	// 	"https://httpbin.org/delay/1",  // may hit global ctx deadline before it even starts
	// 	"https://httpbin.org/delay/1",
	// 	"https://httpbin.org/delay/1",
	// }
	//endregion

	// Option F — everything at once, larger list. Set limit to 1, 4, 16 and
	// compare; check that the order of results still matches the input order.
	urls = []string{
		"https://example.com",
		"https://go.dev",
		"https://httpbin.org/status/200",
		"https://httpbin.org/status/404",
		"https://httpbin.org/status/500",
		"https://httpbin.org/status/503",
		"https://httpbin.org/delay/2",
		"https://httpbin.org/delay/7",
		"https://nope.invalid",
		"http://localhost:1",
		"https://expired.badssl.com",
		"https://self-signed.badssl.com",
		"https://www.cloudflare.com",
		"https://github.com",
	}
	for _, r := range CheckAll(ctx, urls, 2) {
		fmt.Printf("%-40s status=%d latency=%v err=%v\n", r.URL, r.Status, r.Latency, r.Err)
	}
}