[Metrics] Add metric to monitor timeout canceled fragment count

be/src/http/ev_http_server.cpp

@@ -92,7 +92,8 @@ EvHttpServer::~EvHttpServer() {
 
 void EvHttpServer::start() {
     // bind to 
-    CHECK(_bind().ok());
+    auto s = _bind();
+    CHECK(s.ok()) << s.to_string();
     ThreadPoolBuilder("EvHttpServer")
             .set_min_threads(_num_workers)
             .set_max_threads(_num_workers)

be/src/runtime/fragment_mgr.cpp

@@ -53,6 +53,7 @@
 namespace doris {
 
 DEFINE_GAUGE_METRIC_PROTOTYPE_2ARG(plan_fragment_count, MetricUnit::NOUNIT);
+DEFINE_GAUGE_METRIC_PROTOTYPE_2ARG(timeout_canceled_fragment_count, MetricUnit::NOUNIT);
 
 std::string to_load_error_http_path(const std::string& file_name) {
     if (file_name.empty()) {
@@ -381,16 +382,19 @@ FragmentMgr::FragmentMgr(ExecEnv* exec_env)
         : _exec_env(exec_env),
           _fragment_map(),
           _stop_background_threads_latch(1) {
+    _entity = DorisMetrics::instance()->metric_registry()->register_entity("FragmentMgr");
+    INT_UGAUGE_METRIC_REGISTER(_entity, timeout_canceled_fragment_count);
+    REGISTER_HOOK_METRIC(plan_fragment_count, [this]() {
+        std::lock_guard<std::mutex> lock(_lock);
+        return _fragment_map.size();
+    });
+
     CHECK(Thread::create("FragmentMgr", "cancel_timeout_plan_fragment",
                          [this]() {
                              this->cancel_worker();
                          },
                          &_cancel_thread).ok());
 
-    REGISTER_HOOK_METRIC(plan_fragment_count, [this]() {
-        std::lock_guard<std::mutex> lock(_lock);
-        return _fragment_map.size();
-    });
     // TODO(zc): we need a better thread-pool
     // now one user can use all the thread pool, others have no resource.
     ThreadPoolBuilder("FragmentMgrThreadPool")
@@ -523,6 +527,7 @@ void FragmentMgr::cancel_worker() {
                 }
             }
         }
+        timeout_canceled_fragment_count->increment(to_delete.size());
         for (auto& id : to_delete) {
             cancel(id, PPlanFragmentCancelReason::TIMEOUT);
             LOG(INFO) << "FragmentMgr cancel worker going to cancel timeout fragment " << print_id(id);

be/src/runtime/fragment_mgr.h

@@ -33,6 +33,7 @@
 #include "http/rest_monitor_iface.h"
 #include "gutil/ref_counted.h"
 #include "util/countdown_latch.h"
+#include "util/metrics.h"
 #include "util/thread.h"
 
 namespace doris {
@@ -93,6 +94,9 @@ class FragmentMgr : public RestMonitorIface {
     scoped_refptr<Thread> _cancel_thread;
     // every job is a pool
     std::unique_ptr<ThreadPool> _thread_pool;
+
+    std::shared_ptr<MetricEntity> _entity = nullptr;
+    UIntGauge* timeout_canceled_fragment_count = nullptr;
 };
 
 }