Add production hardening: monitoring, backups, caching, and high-traffic resilience

- Recovery middleware (no more crash-on-panic), request metrics/monitoring store + admin API, in-memory TTL caching for read-heavy option/question/ category lookups, gzip + rate-limiting middleware, sanitized chat message encoding, and pg_dump-based database backups with an admin API. - Fix N+1 queries in the discover feed, candidate detail, and match list endpoints by batching selection/photo/user/last-message lookups into single queries instead of one-per-item loops. - Tune DB connection pooling and add HTTP server timeouts + graceful shutdown so the API degrades gracefully instead of falling over under load. - Fix a null (vs empty array) JSON response for candidates/matches without photos or interests, which crashed the web app's swipe card.
2026-09-19 22:17:12 +03:30
parent 15583009e0
commit 769e76313f
29 changed files with 1377 additions and 72 deletions
@@ -0,0 +1,134 @@
+// Package monitoring یک ذخیره‌ساز سبک و درون‌حافظه‌ای برای دیدبانی درخواست‌هاست:
+// آخرین درخواست‌ها (برای دیباگ لحظه‌ای) + آمار تجمیعی (بار، زمان پاسخ، نرخ خطا).
+// عمداً بدون وابستگی بیرونی (بدون Prometheus/Redis) چون در این مقیاس یک حلقه‌ی
+// درون‌حافظه‌ای thread-safe کافیه؛ اگه بعداً مقیاس بزرگ‌تر شد، همین Store رو می‌شه
+// پشت یک exporter واقعی گذاشت.
+package monitoring
+
+import (
+	"sort"
+	"sync"
+	"time"
+)
+
+const bufferSize = 1000
+
+type RequestRecord struct {
+	Time       time.Time `json:"time"`
+	Method     string    `json:"method"`
+	Path       string    `json:"path"`
+	StatusCode int       `json:"status_code"`
+	DurationMs float64   `json:"duration_ms"`
+	ClientIP   string    `json:"client_ip"`
+	UserID     *uint     `json:"user_id,omitempty"`
+}
+
+type Summary struct {
+	TotalRequests   int64     `json:"total_requests"`
+	TotalErrors     int64     `json:"total_errors"`
+	ErrorRatePct    float64   `json:"error_rate_pct"`
+	AvgLatencyMs    float64   `json:"avg_latency_ms"`
+	P95LatencyMs    float64   `json:"p95_latency_ms"`
+	RequestsLastMin int       `json:"requests_last_min"`
+	StartedAt       time.Time `json:"started_at"`
+	UptimeSeconds   float64   `json:"uptime_seconds"`
+}
+
+// Store یک حلقه‌ی ثابت‌اندازه (ring buffer) از آخرین درخواست‌ها به‌همراه
+// شمارنده‌های تجمیعیه. نوشتن/خواندن با RWMutex محافظت می‌شه؛ هزینه‌ی هر
+// درخواست فقط یک قفل کوتاه و بدون I/O روی دیسکه (سریع، مناسب مسیر داغ).
+type Store struct {
+	mu         sync.RWMutex
+	buf        []RequestRecord
+	next       int
+	count      int
+	total      int64
+	totalErr   int64
+	sumLatency float64
+	startedAt  time.Time
+}
+
+func NewStore() *Store {
+	return &Store{
+		buf:       make([]RequestRecord, bufferSize),
+		startedAt: time.Now(),
+	}
+}
+
+func (s *Store) Record(rec RequestRecord) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	s.buf[s.next] = rec
+	s.next = (s.next + 1) % bufferSize
+	if s.count < bufferSize {
+		s.count++
+	}
+	s.total++
+	s.sumLatency += rec.DurationMs
+	if rec.StatusCode >= 400 {
+		s.totalErr++
+	}
+}
+
+// Recent آخرین n درخواست رو برمی‌گردونه (جدیدترین اول).
+func (s *Store) Recent(n int) []RequestRecord {
+	s.mu.RLock()
+	defer s.mu.RUnlock()
+
+	if n > s.count {
+		n = s.count
+	}
+	out := make([]RequestRecord, 0, n)
+	idx := s.next - 1
+	for i := 0; i < n; i++ {
+		if idx < 0 {
+			idx = bufferSize - 1
+		}
+		out = append(out, s.buf[idx])
+		idx--
+	}
+	return out
+}
+
+func (s *Store) Summary() Summary {
+	s.mu.RLock()
+	defer s.mu.RUnlock()
+
+	var errRate, avgLatency float64
+	if s.total > 0 {
+		errRate = float64(s.totalErr) / float64(s.total) * 100
+		avgLatency = s.sumLatency / float64(s.total)
+	}
+
+	cutoff := time.Now().Add(-1 * time.Minute)
+	lastMin := 0
+	latencies := make([]float64, 0, s.count)
+	for i := 0; i < s.count; i++ {
+		r := s.buf[i]
+		if r.Time.After(cutoff) {
+			lastMin++
+		}
+		latencies = append(latencies, r.DurationMs)
+	}
+	sort.Float64s(latencies)
+	p95 := 0.0
+	if len(latencies) > 0 {
+		idx := int(float64(len(latencies)) * 0.95)
+		if idx >= len(latencies) {
+			idx = len(latencies) - 1
+		}
+		p95 = latencies[idx]
+	}
+
+	return Summary{
+		TotalRequests:   s.total,
+		TotalErrors:     s.totalErr,
+		ErrorRatePct:    errRate,
+		AvgLatencyMs:    avgLatency,
+		P95LatencyMs:    p95,
+		RequestsLastMin: lastMin,
+		StartedAt:       s.startedAt,
+		UptimeSeconds:   time.Since(s.startedAt).Seconds(),
+	}
+}
@@ -0,0 +1,74 @@
+package monitoring
+
+import (
+	"testing"
+	"time"
+)
+
+func TestStore_RecordAndSummary(t *testing.T) {
+	s := NewStore()
+
+	s.Record(RequestRecord{Time: time.Now(), Method: "GET", Path: "/x", StatusCode: 200, DurationMs: 10})
+	s.Record(RequestRecord{Time: time.Now(), Method: "GET", Path: "/x", StatusCode: 500, DurationMs: 20})
+
+	sum := s.Summary()
+	if sum.TotalRequests != 2 {
+		t.Fatalf("expected 2 total requests, got %d", sum.TotalRequests)
+	}
+	if sum.TotalErrors != 1 {
+		t.Fatalf("expected 1 error (status>=400), got %d", sum.TotalErrors)
+	}
+	if sum.ErrorRatePct != 50 {
+		t.Fatalf("expected 50%% error rate, got %v", sum.ErrorRatePct)
+	}
+	if sum.AvgLatencyMs != 15 {
+		t.Fatalf("expected avg latency 15ms, got %v", sum.AvgLatencyMs)
+	}
+}
+
+func TestStore_RecentOrder(t *testing.T) {
+	s := NewStore()
+	s.Record(RequestRecord{Time: time.Now(), Path: "/1", StatusCode: 200})
+	s.Record(RequestRecord{Time: time.Now(), Path: "/2", StatusCode: 200})
+	s.Record(RequestRecord{Time: time.Now(), Path: "/3", StatusCode: 200})
+
+	recent := s.Recent(10)
+	if len(recent) != 3 {
+		t.Fatalf("expected 3 records, got %d", len(recent))
+	}
+	// جدیدترین باید اول باشه
+	if recent[0].Path != "/3" || recent[1].Path != "/2" || recent[2].Path != "/1" {
+		t.Fatalf("expected newest-first order, got %+v", recent)
+	}
+}
+
+func TestStore_RecentLimit(t *testing.T) {
+	s := NewStore()
+	for i := 0; i < 5; i++ {
+		s.Record(RequestRecord{Time: time.Now(), Path: "/x", StatusCode: 200})
+	}
+	if got := len(s.Recent(2)); got != 2 {
+		t.Fatalf("expected limit of 2, got %d", got)
+	}
+	if got := len(s.Recent(100)); got != 5 {
+		t.Fatalf("expected all 5 when limit exceeds count, got %d", got)
+	}
+}
+
+func TestStore_RingBufferWraparound(t *testing.T) {
+	s := NewStore()
+	// بیشتر از ظرفیت بافر ثبت می‌کنیم تا رفتار wraparound رو تست کنیم
+	for i := 0; i < bufferSize+50; i++ {
+		s.Record(RequestRecord{Time: time.Now(), Path: "/x", StatusCode: 200, DurationMs: 1})
+	}
+
+	sum := s.Summary()
+	if sum.TotalRequests != int64(bufferSize+50) {
+		t.Fatalf("expected total requests to keep counting past buffer size, got %d", sum.TotalRequests)
+	}
+
+	recent := s.Recent(bufferSize + 100)
+	if len(recent) != bufferSize {
+		t.Fatalf("expected Recent to be capped at bufferSize (%d), got %d", bufferSize, len(recent))
+	}
+}