Upgrade core engine to v3.3.3 LTS

Author: SamNet-dev

.github/workflows/build-engine.yml

@@ -6,11 +6,11 @@ on:
       telemt_commit:
         description: 'Telemt commit hash to build from'
         required: true
-        default: 'e11da6d'
+        default: 'ef7dc2b'
       version_tag:
-        description: 'Version tag (e.g. 3.1.6-e11da6d)'
+        description: 'Version tag (e.g. 3.3.3-ef7dc2b)'
         required: true
-        default: '3.1.6-e11da6d'
+        default: '3.3.3-ef7dc2b'
 
 env:
   REGISTRY: ghcr.io

README.md

@@ -500,20 +500,18 @@ mtproxymax telegram remove              # Remove bot completely
 
 ## 📋 Changelog
 
-### v1.0.0 — Engine v3.1.6
-
-**Engine Upgrade (v3.1.2 → v3.1.6):**
-
-- **Writer Lifecycle Model** — Pool writers now follow Warm → Active → Draining lifecycle with generation-aware safe promotion
-- **Shadow Writers** — Single-endpoint DCs get shadow writers for resilience, with periodic rotation to prevent silent decay
-- **Outage Recovery** — Automatic aggressive reconnect loop when all writers are lost, with backoff and optional quarantine bypass
-- **Adaptive Floor** — Dynamic minimum writer target reduces idle churn while preserving fast recovery under load (enabled by default)
-- **DNS Overrides** — Custom DNS resolution at engine level, hot-reloadable
-- **Upstream Retry + Health** — Upstreams retry 3x with backoff, auto-marked unhealthy after consecutive failures
-- **ME/DC Method Detection** — Engine detects optimal connection method per datacenter
-- **Backpressure Tuning** — Adaptive base/high watermark timeouts for ME routing under load
-- **Metrics Security** — Prometheus endpoint binds to localhost only by default
-- **Per-User Ad-Tag** — Individual ad-tags per user with global fallback
+### v1.0.0 — Engine v3.3.3 LTS
+
+**Engine Upgrade (v3.1.6 → v3.3.3 LTS):**
+
+- **ME NoWait Routing** — Connections no longer block waiting for writers, reducing latency spikes under churn
+- **Async Recovery Failfast** — New send path with async recovery instead of inline blocking, fewer stalls under load
+- **Atomic Secret Snapshots** — Lock-free secret reads on every connection for higher throughput at scale
+- **HybridAsyncPersistent** — New ME route mode keeps routing even when writers are temporarily unavailable
+- **Upstream Connection Budget** — Limits concurrent upstream connection attempts, prevents connection storms
+- **IP Limit Rework** — New IP tracking system with unique IP always in metrics
+- **ME Pool Init Fixes** — More reliable pool initialization with retry logic
+- **ME-to-DC Fallback** — Graceful fallback when preferred DC method is unavailable
 
 ### v1.0.0 — Per-User Limits + Telegram Bot
 

mtproxymax.sh

@@ -22,8 +22,8 @@ UPSTREAMS_FILE="${INSTALL_DIR}/upstreams.conf"
 BACKUP_DIR="${INSTALL_DIR}/backups"
 CONTAINER_NAME="mtproxymax"
 DOCKER_IMAGE_BASE="mtproxymax-telemt"
-TELEMT_MIN_VERSION="3.1.6"
-TELEMT_COMMIT="e11da6d"  # Pinned: v3.1.6 — Writer lifecycle, shadow writers, adaptive floor, upstream health, outage recovery
+TELEMT_MIN_VERSION="3.3.3"
+TELEMT_COMMIT="ef7dc2b"  # Pinned: v3.3.3 LTS — NoWait routing, atomic secrets, async recovery, upstream budget, perf improvements
 GITHUB_REPO="SamNet-dev/MTProxyMax"
 REGISTRY_IMAGE="ghcr.io/samnet-dev/mtproxymax-telemt"