diff --git a/tools/load_tester.cpp b/tools/load_tester.cpp
index c5f7187..722bf14 100644
--- a/tools/load_tester.cpp
+++ b/tools/load_tester.cpp
@@ -23,6 +23,10 @@
 
 #include <llhttp.h>
 
+#ifndef __has_feature
+#define __has_feature(x) 0
+#endif
+
 // Use shared perfetto categories
 #include "../src/perfetto_categories.hpp"
 
@@ -139,6 +143,7 @@ sem_t connectionLimit;
 
 // Shutdown mechanism
 std::atomic<bool> g_shutdown{false};
+std::atomic<int> g_connect_threads{0};
 
 void signal_handler(int sig) {
   if (sig == SIGTERM || sig == SIGINT) {
@@ -167,13 +172,44 @@ struct Connection {
   uint64_t requestId = 0;
   void initRequest() {
     requestId = nextRequestId.fetch_add(1, std::memory_order_relaxed);
-    int len = snprintf(buf, sizeof(buf),
-                       "GET /ok HTTP/1.1\r\nX-Request-Id: %" PRIu64 "\r\n\r\n",
-                       requestId);
-    if (len == -1 || len > int(sizeof(buf))) {
+
+    // Fast manual string construction - avoid snprintf overhead
+    const char *prefix = "GET /ok HTTP/1.1\r\nX-Request-Id: ";
+    const char *suffix = "\r\n\r\n";
+
+    // Copy prefix
+    char *p = buf;
+    const char *src = prefix;
+    while (*src)
+      *p++ = *src++;
+
+    // Convert requestId to string directly
+    if (requestId == 0) {
+      *p++ = '0';
+    } else {
+      char digits[20]; // uint64_t max is 20 digits
+      int digit_count = 0;
+      uint64_t temp = requestId;
+      while (temp > 0) {
+        digits[digit_count++] = '0' + (temp % 10);
+        temp /= 10;
+      }
+      // Copy digits in reverse order
+      for (int i = digit_count - 1; i >= 0; i--) {
+        *p++ = digits[i];
+      }
+    }
+
+    // Copy suffix
+    src = suffix;
+    while (*src)
+      *p++ = *src++;
+
+    size_t len = p - buf;
+    if (len >= sizeof(buf)) {
       abort();
     }
-    request = std::string_view{buf, size_t(len)};
+    request = std::string_view{buf, len};
   }
 
   Connection(int fd, int64_t id) : fd(fd), id(id) {
@@ -534,6 +570,7 @@ int main(int argc, char *argv[]) {
     abort();
   }
   std::atomic<int64_t> connectionId{0};
+  g_connect_threads.store(g_config.connect_threads, std::memory_order_relaxed);
 
   std::vector<std::thread> threads;
 
@@ -541,7 +578,7 @@ int main(int argc, char *argv[]) {
     threads.emplace_back([epollfd, i]() {
       pthread_setname_np(pthread_self(),
                          ("network-" + std::to_string(i)).c_str());
-      while (!g_shutdown.load(std::memory_order_relaxed)) {
+      while (g_connect_threads.load() != 0) {
         struct epoll_event events[256]; // Use a reasonable max size
         int batch_size = std::min(int(sizeof(events) / sizeof(events[0])),
                                   g_config.event_batch_size);
@@ -660,6 +697,7 @@ int main(int argc, char *argv[]) {
           continue;
         }
       }
+      g_connect_threads.fetch_sub(1);
     });
   }