- cache: fix LRU race — check expiry before MoveToFront within same lock critical section; previously a stale entry could be promoted before expiry was verified - cache: drop stale ETag/Content-Md5/Digest headers when body is rewritten by refreshCachedResponseMeta to avoid misrepresenting the new body to clients - ratelimit: cap buckets map at 100k entries with stale-eviction fallback and shared overflow bucket to prevent unbounded memory growth under high-cardinality traffic - ratelimit: fix clientRateLimitKey comment — credentials are tried second (before IP), not as a "last resort" Co-Authored-By: Virgil <virgil@lethean.io>
276 lines
6.7 KiB
Go
276 lines
6.7 KiB
Go
// SPDX-License-Identifier: EUPL-1.2
|
|
|
|
package api
|
|
|
|
import (
|
|
"crypto/sha256"
|
|
"encoding/hex"
|
|
"math"
|
|
"net/http"
|
|
"strconv"
|
|
"strings"
|
|
"sync"
|
|
"time"
|
|
|
|
"github.com/gin-gonic/gin"
|
|
)
|
|
|
|
const (
|
|
rateLimitCleanupInterval = time.Minute
|
|
rateLimitStaleAfter = 10 * time.Minute
|
|
|
|
// rateLimitMaxBuckets caps the total number of tracked keys to prevent
|
|
// unbounded memory growth under high-cardinality traffic (e.g. scanning
|
|
// bots cycling random IPs). When the cap is reached, new keys that cannot
|
|
// evict a stale bucket are routed to a shared overflow bucket so requests
|
|
// are still rate-limited rather than bypassing the limiter entirely.
|
|
rateLimitMaxBuckets = 100_000
|
|
rateLimitOverflowKey = "__overflow__"
|
|
)
|
|
|
|
type rateLimitStore struct {
|
|
mu sync.Mutex
|
|
buckets map[string]*rateLimitBucket
|
|
limit int
|
|
lastSweep time.Time
|
|
}
|
|
|
|
type rateLimitBucket struct {
|
|
mu sync.Mutex
|
|
tokens float64
|
|
last time.Time
|
|
lastSeen time.Time
|
|
}
|
|
|
|
type rateLimitDecision struct {
|
|
allowed bool
|
|
retryAfter time.Duration
|
|
limit int
|
|
remaining int
|
|
resetAt time.Time
|
|
}
|
|
|
|
func newRateLimitStore(limit int) *rateLimitStore {
|
|
now := time.Now()
|
|
return &rateLimitStore{
|
|
buckets: make(map[string]*rateLimitBucket),
|
|
limit: limit,
|
|
lastSweep: now,
|
|
}
|
|
}
|
|
|
|
func (s *rateLimitStore) allow(key string) rateLimitDecision {
|
|
now := time.Now()
|
|
|
|
s.mu.Lock()
|
|
bucket, ok := s.buckets[key]
|
|
if !ok || now.Sub(bucket.lastSeen) > rateLimitStaleAfter {
|
|
// Enforce the bucket cap before inserting a new entry. First try to
|
|
// evict a single stale entry; if none exists and the map is full,
|
|
// route the request to the shared overflow bucket so it is still
|
|
// rate-limited rather than bypassing the limiter.
|
|
if !ok && len(s.buckets) >= rateLimitMaxBuckets {
|
|
evicted := false
|
|
for k, candidate := range s.buckets {
|
|
if now.Sub(candidate.lastSeen) > rateLimitStaleAfter {
|
|
delete(s.buckets, k)
|
|
evicted = true
|
|
break
|
|
}
|
|
}
|
|
if !evicted {
|
|
// Cap reached and no stale entry to evict: use overflow bucket.
|
|
key = rateLimitOverflowKey
|
|
if ob, exists := s.buckets[key]; exists {
|
|
bucket = ob
|
|
ok = true
|
|
}
|
|
}
|
|
}
|
|
|
|
if !ok {
|
|
bucket = &rateLimitBucket{
|
|
tokens: float64(s.limit),
|
|
last: now,
|
|
lastSeen: now,
|
|
}
|
|
s.buckets[key] = bucket
|
|
} else {
|
|
bucket.lastSeen = now
|
|
}
|
|
} else {
|
|
bucket.lastSeen = now
|
|
}
|
|
|
|
if now.Sub(s.lastSweep) >= rateLimitCleanupInterval {
|
|
for k, candidate := range s.buckets {
|
|
if now.Sub(candidate.lastSeen) > rateLimitStaleAfter {
|
|
delete(s.buckets, k)
|
|
}
|
|
}
|
|
s.lastSweep = now
|
|
}
|
|
s.mu.Unlock()
|
|
|
|
bucket.mu.Lock()
|
|
defer bucket.mu.Unlock()
|
|
|
|
elapsed := now.Sub(bucket.last)
|
|
if elapsed > 0 {
|
|
refill := elapsed.Seconds() * float64(s.limit)
|
|
if bucket.tokens+refill > float64(s.limit) {
|
|
bucket.tokens = float64(s.limit)
|
|
} else {
|
|
bucket.tokens += refill
|
|
}
|
|
bucket.last = now
|
|
}
|
|
|
|
if bucket.tokens >= 1 {
|
|
bucket.tokens--
|
|
return rateLimitDecision{
|
|
allowed: true,
|
|
limit: s.limit,
|
|
remaining: int(math.Floor(bucket.tokens)),
|
|
resetAt: now.Add(timeUntilFull(bucket.tokens, s.limit)),
|
|
}
|
|
}
|
|
|
|
deficit := 1 - bucket.tokens
|
|
wait := time.Duration(deficit / float64(s.limit) * float64(time.Second))
|
|
if wait <= 0 {
|
|
wait = time.Second / time.Duration(s.limit)
|
|
if wait <= 0 {
|
|
wait = time.Second
|
|
}
|
|
}
|
|
|
|
return rateLimitDecision{
|
|
allowed: false,
|
|
retryAfter: wait,
|
|
limit: s.limit,
|
|
remaining: 0,
|
|
resetAt: now.Add(wait),
|
|
}
|
|
}
|
|
|
|
func rateLimitMiddleware(limit int) gin.HandlerFunc {
|
|
if limit <= 0 {
|
|
return func(c *gin.Context) {
|
|
c.Next()
|
|
}
|
|
}
|
|
|
|
store := newRateLimitStore(limit)
|
|
|
|
return func(c *gin.Context) {
|
|
key := clientRateLimitKey(c)
|
|
decision := store.allow(key)
|
|
if !decision.allowed {
|
|
secs := int(decision.retryAfter / time.Second)
|
|
if decision.retryAfter%time.Second != 0 {
|
|
secs++
|
|
}
|
|
if secs < 1 {
|
|
secs = 1
|
|
}
|
|
setRateLimitHeaders(c, decision.limit, decision.remaining, decision.resetAt)
|
|
c.Header("Retry-After", strconv.Itoa(secs))
|
|
c.AbortWithStatusJSON(http.StatusTooManyRequests, Fail(
|
|
"rate_limit_exceeded",
|
|
"Too many requests",
|
|
))
|
|
return
|
|
}
|
|
|
|
setRateLimitHeaders(c, decision.limit, decision.remaining, decision.resetAt)
|
|
c.Next()
|
|
}
|
|
}
|
|
|
|
func setRateLimitHeaders(c *gin.Context, limit, remaining int, resetAt time.Time) {
|
|
if limit > 0 {
|
|
c.Header("X-RateLimit-Limit", strconv.Itoa(limit))
|
|
}
|
|
if remaining < 0 {
|
|
remaining = 0
|
|
}
|
|
c.Header("X-RateLimit-Remaining", strconv.Itoa(remaining))
|
|
if !resetAt.IsZero() {
|
|
reset := resetAt.Unix()
|
|
if reset <= time.Now().Unix() {
|
|
reset = time.Now().Add(time.Second).Unix()
|
|
}
|
|
c.Header("X-RateLimit-Reset", strconv.FormatInt(reset, 10))
|
|
}
|
|
}
|
|
|
|
func timeUntilFull(tokens float64, limit int) time.Duration {
|
|
if limit <= 0 {
|
|
return 0
|
|
}
|
|
missing := float64(limit) - tokens
|
|
if missing <= 0 {
|
|
return 0
|
|
}
|
|
seconds := missing / float64(limit)
|
|
if seconds <= 0 {
|
|
return 0
|
|
}
|
|
return time.Duration(math.Ceil(seconds * float64(time.Second)))
|
|
}
|
|
|
|
// clientRateLimitKey derives a bucket key for the request. It prefers a
|
|
// validated principal placed in context by auth middleware, then falls back to
|
|
// raw credential headers (X-API-Key or Bearer token, hashed with SHA-256 so
|
|
// secrets are never stored in the bucket map), and finally falls back to the
|
|
// client IP when no credentials are present.
|
|
func clientRateLimitKey(c *gin.Context) string {
|
|
// Prefer a validated principal placed in context by auth middleware.
|
|
if principal, ok := c.Get("principal"); ok && principal != nil {
|
|
if s, ok := principal.(string); ok && s != "" {
|
|
return "principal:" + s
|
|
}
|
|
}
|
|
if userID, ok := c.Get("userID"); ok && userID != nil {
|
|
if s, ok := userID.(string); ok && s != "" {
|
|
return "user:" + s
|
|
}
|
|
}
|
|
|
|
// Fall back to credential headers before the IP so that different API
|
|
// keys coming from the same NAT address are bucketed independently. The
|
|
// raw secret is never stored — it is hashed with SHA-256 first.
|
|
if apiKey := strings.TrimSpace(c.GetHeader("X-API-Key")); apiKey != "" {
|
|
h := sha256.Sum256([]byte(apiKey))
|
|
return "cred:sha256:" + hex.EncodeToString(h[:])
|
|
}
|
|
if bearer := bearerTokenFromHeader(c.GetHeader("Authorization")); bearer != "" {
|
|
h := sha256.Sum256([]byte(bearer))
|
|
return "cred:sha256:" + hex.EncodeToString(h[:])
|
|
}
|
|
|
|
// Last resort: fall back to IP address.
|
|
if ip := c.ClientIP(); ip != "" {
|
|
return "ip:" + ip
|
|
}
|
|
if c.Request != nil && c.Request.RemoteAddr != "" {
|
|
return "ip:" + c.Request.RemoteAddr
|
|
}
|
|
|
|
return "ip:unknown"
|
|
}
|
|
|
|
func bearerTokenFromHeader(header string) string {
|
|
header = strings.TrimSpace(header)
|
|
if header == "" {
|
|
return ""
|
|
}
|
|
|
|
parts := strings.SplitN(header, " ", 2)
|
|
if len(parts) != 2 || !strings.EqualFold(parts[0], "Bearer") {
|
|
return ""
|
|
}
|
|
|
|
return strings.TrimSpace(parts[1])
|
|
}
|