Improve telego client robustness and retries

Add a createRobustFastHTTPClient helper to configure fasthttp.Client with better timeouts, connection limits, retries and optional SOCKS5 proxy dialing. Validate and sanitize proxy and API server URLs instead of returning early on invalid values, and build telego.Bot options dynamically. Reduce long-polling timeout to detect connection issues faster and adjust update retrieval comments. Implement exponential-backoff retry logic for SendMessage calls to handle transient connection/timeouts and improve delivery reliability; also reduce inter-message delay for better throughput.

3 files changed, 117 insertions, 54 deletions

diff --git a/web/service/inbound.go b/web/service/inbound.go
index ec51bc27..101c79d9 100644
--- a/web/service/inbound.go
+++ b/web/service/inbound.go
@@ -2141,25 +2141,25 @@ func (s *InboundService) GetInboundClientIps(clientEmail string) (string, error)
 	if err != nil {
 		return "", err
 	}
-	
+
 	if InboundClientIps.Ips == "" {
 		return "", nil
 	}
-	
+
 	// Try to parse as new format (with timestamps)
 	type IPWithTimestamp struct {
 		IP        string `json:"ip"`
 		Timestamp int64  `json:"timestamp"`
 	}
-	
+
 	var ipsWithTime []IPWithTimestamp
 	err = json.Unmarshal([]byte(InboundClientIps.Ips), &ipsWithTime)
-	
+
 	// If successfully parsed as new format, return with timestamps
 	if err == nil && len(ipsWithTime) > 0 {
 		return InboundClientIps.Ips, nil
 	}
-	
+
 	// Otherwise, assume it's old format (simple string array)
 	// Try to parse as simple array and convert to new format
 	var oldIps []string
@@ -2176,7 +2176,7 @@ func (s *InboundService) GetInboundClientIps(clientEmail string) (string, error)
 		result, _ := json.Marshal(newIpsWithTime)
 		return string(result), nil
 	}
-	
+
 	// Return as-is if parsing fails
 	return InboundClientIps.Ips, nil
 }
diff --git a/web/service/setting.go b/web/service/setting.go
index 1e95c2df..33012c39 100644
--- a/web/service/setting.go
+++ b/web/service/setting.go
@@ -5,7 +5,7 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-    "net"
+	"net"
 	"reflect"
 	"strconv"
 	"strings"
@@ -719,25 +719,25 @@ func (s *SettingService) GetDefaultXrayConfig() (any, error) {
 }
 
 func extractHostname(host string) string {
-    h, _, err := net.SplitHostPort(host)
-    // Err is not nil means host does not contain port
-    if err != nil {
-        h = host
-    }
-
-    ip := net.ParseIP(h)
-    // If it's not an IP, return as is
-    if ip == nil {
-        return h
-    }
-
-    // If it's an IPv4, return as is
-    if ip.To4() != nil {
-        return h
-    }
-
-    // IPv6 needs bracketing
-    return "[" + h + "]"
+	h, _, err := net.SplitHostPort(host)
+	// Err is not nil means host does not contain port
+	if err != nil {
+		h = host
+	}
+
+	ip := net.ParseIP(h)
+	// If it's not an IP, return as is
+	if ip == nil {
+		return h
+	}
+
+	// If it's an IPv4, return as is
+	if ip.To4() != nil {
+		return h
+	}
+
+	// IPv6 needs bracketing
+	return "[" + h + "]"
 }
 
 func (s *SettingService) GetDefaultSettings(host string) (any, error) {
diff --git a/web/service/tgbot.go b/web/service/tgbot.go
index 15f5dbc2..3ff80b40 100644
--- a/web/service/tgbot.go
+++ b/web/service/tgbot.go
@@ -272,41 +272,78 @@ func (t *Tgbot) Start(i18nFS embed.FS) error {
 	return nil
 }
 
-// NewBot creates a new Telegram bot instance with optional proxy and API server settings.
-func (t *Tgbot) NewBot(token string, proxyUrl string, apiServerUrl string) (*telego.Bot, error) {
-	if proxyUrl == "" && apiServerUrl == "" {
-		return telego.NewBot(token)
+// createRobustFastHTTPClient creates a fasthttp.Client with proper connection handling
+func (t *Tgbot) createRobustFastHTTPClient(proxyUrl string) *fasthttp.Client {
+	client := &fasthttp.Client{
+		// Connection timeouts
+		ReadTimeout:                   30 * time.Second,
+		WriteTimeout:                  30 * time.Second,
+		MaxIdleConnDuration:           60 * time.Second,
+		MaxConnDuration:               0, // unlimited, but controlled by MaxIdleConnDuration
+		MaxIdemponentCallAttempts:     3,
+		ReadBufferSize:                4096,
+		WriteBufferSize:               4096,
+		MaxConnsPerHost:               100,
+		MaxConnWaitTimeout:            10 * time.Second,
+		DisableHeaderNamesNormalizing: false,
+		DisablePathNormalizing:        false,
+		// Retry on connection errors
+		RetryIf: func(request *fasthttp.Request) bool {
+			// Retry on connection errors for GET requests
+			return string(request.Header.Method()) == "GET" || string(request.Header.Method()) == "POST"
+		},
+	}
+
+	// Set proxy if provided
+	if proxyUrl != "" {
+		client.Dial = fasthttpproxy.FasthttpSocksDialer(proxyUrl)
 	}
 
+	return client
+}
+
+// NewBot creates a new Telegram bot instance with optional proxy and API server settings.
+func (t *Tgbot) NewBot(token string, proxyUrl string, apiServerUrl string) (*telego.Bot, error) {
+	// Validate proxy URL if provided
 	if proxyUrl != "" {
 		if !strings.HasPrefix(proxyUrl, "socks5://") {
-			logger.Warning("Invalid socks5 URL, using default")
-			return telego.NewBot(token)
+			logger.Warning("Invalid socks5 URL, ignoring proxy")
+			proxyUrl = "" // Clear invalid proxy
+		} else {
+			_, err := url.Parse(proxyUrl)
+			if err != nil {
+				logger.Warningf("Can't parse proxy URL, ignoring proxy: %v", err)
+				proxyUrl = ""
+			}
 		}
+	}
 
-		_, err := url.Parse(proxyUrl)
-		if err != nil {
-			logger.Warningf("Can't parse proxy URL, using default instance for tgbot: %v", err)
-			return telego.NewBot(token)
+	// Validate API server URL if provided
+	if apiServerUrl != "" {
+		if !strings.HasPrefix(apiServerUrl, "http") {
+			logger.Warning("Invalid http(s) URL for API server, using default")
+			apiServerUrl = ""
+		} else {
+			_, err := url.Parse(apiServerUrl)
+			if err != nil {
+				logger.Warningf("Can't parse API server URL, using default: %v", err)
+				apiServerUrl = ""
+			}
 		}
-
-		return telego.NewBot(token, telego.WithFastHTTPClient(&fasthttp.Client{
-			Dial: fasthttpproxy.FasthttpSocksDialer(proxyUrl),
-		}))
 	}
 
-	if !strings.HasPrefix(apiServerUrl, "http") {
-		logger.Warning("Invalid http(s) URL, using default")
-		return telego.NewBot(token)
-	}
+	// Create robust fasthttp client
+	client := t.createRobustFastHTTPClient(proxyUrl)
 
-	_, err := url.Parse(apiServerUrl)
-	if err != nil {
-		logger.Warningf("Can't parse API server URL, using default instance for tgbot: %v", err)
-		return telego.NewBot(token)
+	// Build bot options
+	var options []telego.BotOption
+	options = append(options, telego.WithFastHTTPClient(client))
+
+	if apiServerUrl != "" {
+		options = append(options, telego.WithAPIServer(apiServerUrl))
 	}
 
-	return telego.NewBot(token, telego.WithAPIServer(apiServerUrl))
+	return telego.NewBot(token, options...)
 }
 
 // IsRunning checks if the Telegram bot is currently running.
@@ -390,7 +427,7 @@ func (t *Tgbot) decodeQuery(query string) (string, error) {
 // OnReceive starts the message receiving loop for the Telegram bot.
 func (t *Tgbot) OnReceive() {
 	params := telego.GetUpdatesParams{
-		Timeout: 30, // Increased timeout to reduce API calls
+		Timeout: 20, // Reduced timeout to detect connection issues faster
 	}
 	// Strict singleton: never start a second long-polling loop.
 	tgBotMutex.Lock()
@@ -408,7 +445,7 @@ func (t *Tgbot) OnReceive() {
 	botWG.Add(1)
 	tgBotMutex.Unlock()
 
-	// Get updates channel using the context.
+	// Get updates channel using the context with shorter timeout for better error recovery
 	updates, _ := bot.UpdatesViaLongPolling(ctx, &params)
 	go func() {
 		defer botWG.Done()
@@ -2247,10 +2284,36 @@ func (t *Tgbot) SendMsgToTgbot(chatId int64, msg string, replyMarkup ...telego.R
 		if len(replyMarkup) > 0 && n == (len(allMessages)-1) {
 			params.ReplyMarkup = replyMarkup[0]
 		}
-		_, err := bot.SendMessage(context.Background(), &params)
-		if err != nil {
-			logger.Warning("Error sending telegram message :", err)
+
+		// Retry logic with exponential backoff for connection errors
+		maxRetries := 3
+		for attempt := range maxRetries {
+			ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+			_, err := bot.SendMessage(ctx, &params)
+			cancel()
+
+			if err == nil {
+				break // Success
+			}
+
+			// Check if error is a connection error
+			errStr := err.Error()
+			isConnectionError := strings.Contains(errStr, "connection") ||
+				strings.Contains(errStr, "timeout") ||
+				strings.Contains(errStr, "closed")
+
+			if isConnectionError && attempt < maxRetries-1 {
+				// Exponential backoff: 1s, 2s, 4s
+				backoff := time.Duration(1<<uint(attempt)) * time.Second
+				logger.Warningf("Connection error sending telegram message (attempt %d/%d), retrying in %v: %v",
+					attempt+1, maxRetries, backoff, err)
+				time.Sleep(backoff)
+			} else {
+				logger.Warning("Error sending telegram message:", err)
+				break
+			}
 		}
+
 		// Reduced delay to improve performance (only needed for rate limiting)
 		if n < len(allMessages)-1 { // Only delay between messages, not after the last one
 			time.Sleep(100 * time.Millisecond)