diff --git a/ROMFS/px4fmu_common/init.d-posix/rcS b/ROMFS/px4fmu_common/init.d-posix/rcS
index 340cc019011..fb84d0426ac 100644
--- a/ROMFS/px4fmu_common/init.d-posix/rcS
+++ b/ROMFS/px4fmu_common/init.d-posix/rcS
@@ -216,6 +216,7 @@ if ! replay tryapplyparams
 then
 	simulator start -c $simulator_tcp_port
 fi
+load_mon start
 battery_simulator start
 tone_alarm start
 rc_update start
diff --git a/boards/px4/sitl/default.cmake b/boards/px4/sitl/default.cmake
index ea4664dcd70..dc58fac57b7 100644
--- a/boards/px4/sitl/default.cmake
+++ b/boards/px4/sitl/default.cmake
@@ -36,7 +36,7 @@ px4_add_board(
 		fw_pos_control_l1
 		land_detector
 		landing_target_estimator
-		#load_mon
+		load_mon
 		local_position_estimator
 		logger
 		mavlink
diff --git a/boards/px4/sitl/nolockstep.cmake b/boards/px4/sitl/nolockstep.cmake
index 0f38f318d47..8c0c963b7ca 100644
--- a/boards/px4/sitl/nolockstep.cmake
+++ b/boards/px4/sitl/nolockstep.cmake
@@ -36,7 +36,7 @@ px4_add_board(
 		fw_pos_control_l1
 		land_detector
 		landing_target_estimator
-		#load_mon
+		load_mon
 		local_position_estimator
 		logger
 		mavlink
diff --git a/boards/px4/sitl/rtps.cmake b/boards/px4/sitl/rtps.cmake
index f6ef8912670..fa97dbcf2af 100644
--- a/boards/px4/sitl/rtps.cmake
+++ b/boards/px4/sitl/rtps.cmake
@@ -36,7 +36,7 @@ px4_add_board(
 		fw_pos_control_l1
 		land_detector
 		landing_target_estimator
-		#load_mon
+		load_mon
 		local_position_estimator
 		logger
 		mavlink
diff --git a/boards/px4/sitl/test.cmake b/boards/px4/sitl/test.cmake
index b33fed50e46..6d00a84e5ef 100644
--- a/boards/px4/sitl/test.cmake
+++ b/boards/px4/sitl/test.cmake
@@ -35,7 +35,7 @@ px4_add_board(
 		fw_pos_control_l1
 		land_detector
 		landing_target_estimator
-		#load_mon
+		load_mon
 		local_position_estimator
 		logger
 		mavlink
diff --git a/src/modules/load_mon/LoadMon.cpp b/src/modules/load_mon/LoadMon.cpp
index d87c61d6d35..ea22b60f59a 100644
--- a/src/modules/load_mon/LoadMon.cpp
+++ b/src/modules/load_mon/LoadMon.cpp
@@ -86,6 +86,17 @@ void LoadMon::start()
 
 void LoadMon::Run()
 {
+#if defined (__PX4_LINUX)
+
+	if (_proc_fd == nullptr) {	// init fd
+		_proc_fd = fopen("/proc/meminfo", "r");
+
+		if (_proc_fd == nullptr) {
+			PX4_ERR("Failed to open /proc/meminfo");
+		}
+	}
+
+#endif
 	perf_begin(_cycle_perf);
 
 	cpuload();
@@ -100,6 +111,9 @@ void LoadMon::Run()
 
 	if (should_exit()) {
 		ScheduleClear();
+#if defined (__PX4_LINUX)
+		fclose(_proc_fd);
+#endif
 		exit_and_cleanup();
 	}
 
@@ -142,17 +156,74 @@ void LoadMon::cpuload()
 	const float interval_idletime = total_runtime - _last_idle_time;
 #endif
 
-	// get ram usage
-	struct mallinfo mem = mallinfo();
-	float ram_usage = (float)mem.uordblks / mem.arena;
-
 	cpuload_s cpuload{};
 #if defined(__PX4_LINUX)
+	/* following calculation is based on free(1)
+	 * https://gitlab.com/procps-ng/procps/-/blob/master/proc/sysinfo.c */
+	char line[256];
+	int32_t kb_main_total = -1;
+	int32_t kb_main_free = -1;
+	int32_t kb_page_cache = -1;
+	int32_t kb_slab_reclaimable = -1;
+	int32_t kb_main_buffers = -1;
+	int parsedCount = 0;
+
+	if (_proc_fd != nullptr) {
+		while (fgets(line, sizeof(line), _proc_fd)) {
+			if (sscanf(line, "MemTotal: %d kB", &kb_main_total) == 1) {
+				++parsedCount;
+				continue;
+			}
+
+			if (sscanf(line, "MemFree: %d kB", &kb_main_free) == 1) {
+				++parsedCount;
+				continue;
+			}
+
+			if (sscanf(line, "Cached: %d kB", &kb_page_cache) == 1) {
+				++parsedCount;
+				continue;
+			}
+
+			if (sscanf(line, "SReclaimable: %d kB", &kb_slab_reclaimable) == 1) {
+				++parsedCount;
+				continue;
+			}
+
+			if (sscanf(line, "Buffers: %d kB", &kb_main_buffers) == 1) {
+				++parsedCount;
+				continue;
+			}
+		}
+
+		fseek(_proc_fd, 0, SEEK_END);
+
+		if (parsedCount == 5) {
+			int32_t kb_main_cached = kb_page_cache + kb_slab_reclaimable;
+			int32_t mem_used = kb_main_total - kb_main_free - kb_main_cached - kb_main_buffers;
+
+			if (mem_used < 0) {
+				mem_used = kb_main_total - kb_main_free;
+			}
+
+			cpuload.ram_usage = (float)mem_used / kb_main_total;
+
+		} else {
+			PX4_ERR("Could not parse /proc/meminfo");
+			cpuload.ram_usage = -1;
+		}
+
+	} else {
+		cpuload.ram_usage = -1;
+	}
+
 	cpuload.load = interval_spent_time / interval;
 #elif defined(__PX4_NUTTX)
+	// get ram usage
+	struct mallinfo mem = mallinfo();
+	cpuload.ram_usage = (float)mem.uordblks / mem.arena;
 	cpuload.load = 1.f - interval_idletime / interval;
 #endif
-	cpuload.ram_usage = ram_usage;
 	cpuload.timestamp = hrt_absolute_time();
 
 	_cpuload_pub.publish(cpuload);
diff --git a/src/modules/load_mon/LoadMon.hpp b/src/modules/load_mon/LoadMon.hpp
index d1d66153b73..afdbb23d2ba 100644
--- a/src/modules/load_mon/LoadMon.hpp
+++ b/src/modules/load_mon/LoadMon.hpp
@@ -33,7 +33,9 @@
 
 #pragma once
 
+#if defined(__PX4_NUTTX)
 #include <malloc.h>
+#endif
 #include <drivers/drv_hrt.h>
 #include <lib/perf/perf_counter.h>
 #include <px4_platform_common/px4_config.h>
@@ -48,7 +50,6 @@
 
 #if defined(__PX4_LINUX)
 #include <sys/times.h>
-#include <malloc.h>
 #endif
 
 namespace load_mon
@@ -92,6 +93,7 @@ private:
 	uORB::Publication<cpuload_s> _cpuload_pub {ORB_ID(cpuload)};
 
 #if defined(__PX4_LINUX)
+	FILE *_proc_fd = nullptr;
 	/* calculate usage directly from clock ticks on Linux */
 	clock_t _last_total_time_stamp{};
 	clock_t _last_spent_time_stamp{};