fix: monitor worker non-blocking + test fixes
- core/monitor.py: add fast metrics script (0.3s vs 1.4s sleep), collect() now defaults to fast=True for responsive UI - ui/workers.py: MonitorWorker changed from loop mode to single-shot (QTimer re-arms next cycle on finished signal, avoids paramiko blocking in infinite loop) - ui/widgets.py: remove debug print from _kick_one_sample - test_monitor_nonblock.py: fix sample counting by wrapping _on_sample instead of disconnecting signals (old approach missed new workers) - test_process_monitor.py: assertions account for MAX_RENDER_ROWS=200, fix MEM% column index (4 not 3) in sort test
This commit is contained in:
Your Name
+69
-6
@@ -13,13 +13,16 @@ from .ssh_client import SSHConnection
|
||||
class SystemMonitor:
|
||||
"""远程主机的资源监控器（数据全部从 SSH 通道采集，不依赖 agent）"""
|
||||
|
||||
# 一次性获取所有指标的脚本（Linux）
|
||||
# 慢脚本：CPU 差分需要 sleep 1（约 1.4 秒）
|
||||
_LINUX_METRICS_SCRIPT = r"""
|
||||
echo "===CPU==="
|
||||
# 第一次采样 1 秒间隔，用来计算差值
|
||||
read cpu_user cpu_nice cpu_system cpu_idle cpu_iowait cpu_irq cpu_softirq cpu_steal < /proc/stat
|
||||
# 第一次采样 1 秒间隔，用来计算差分
|
||||
awk 'NR==1 {gsub("cpu ",""); print}' /proc/stat > /tmp/.cpu1.$$
|
||||
sleep 1
|
||||
read cpu_user2 cpu_nice2 cpu_system2 cpu_idle2 cpu_iowait2 cpu_irq2 cpu_softirq2 cpu_steal2 < /proc/stat
|
||||
awk 'NR==1 {gsub("cpu ",""); print}' /proc/stat > /tmp/.cpu2.$$
|
||||
read cpu_user cpu_nice cpu_system cpu_idle cpu_iowait cpu_irq cpu_softirq cpu_steal cpu_guest cpu_gnice < /tmp/.cpu1.$$
|
||||
read cpu_user2 cpu_nice2 cpu_system2 cpu_idle2 cpu_iowait2 cpu_irq2 cpu_softirq2 cpu_steal2 cpu_guest2 cpu_gnice2 < /tmp/.cpu2.$$
|
||||
rm -f /tmp/.cpu1.$$ /tmp/.cpu2.$$
|
||||
total1=$((cpu_user+cpu_nice+cpu_system+cpu_idle+cpu_iowait+cpu_irq+cpu_softirq+cpu_steal))
|
||||
total2=$((cpu_user2+cpu_nice2+cpu_system2+cpu_idle2+cpu_iowait2+cpu_irq2+cpu_softirq2+cpu_steal2))
|
||||
idle1=$cpu_idle; idle2=$cpu_idle2
|
||||
@@ -68,6 +71,65 @@ echo "===HOST==="
|
||||
echo "HOSTNAME=$(hostname)"
|
||||
echo "KERNEL=$(uname -r)"
|
||||
echo "OS=$(. /etc/os-release 2>/dev/null && echo "$PRETTY_NAME" || uname -s)"
|
||||
"""
|
||||
|
||||
# 快脚本：不含 sleep 1，但 CPU 数字用即时 busy%（基于 /proc/stat 当前总样本，
|
||||
# 加上 -1 让服务端快速 grep 两次间隔 0.3s 算差分，~0.4s 总耗时）
|
||||
# 用 usleep 微秒精度，0.3s 差分窗口 + awk 算 CPU%
|
||||
_FAST_METRICS_SCRIPT = r"""
|
||||
echo "===CPU==="
|
||||
# 快速差分：0.3 秒间隔（避免 1.5s 阻塞）。误差 ±5% 但响应快
|
||||
# awk 解析 /proc/stat 第一行（cpu 总览），两次采样做差分
|
||||
awk 'NR==1 {gsub("cpu ",""); print}' /proc/stat > /tmp/.cpu1.$$
|
||||
sleep 0.3 2>/dev/null || sleep 1
|
||||
awk 'NR==1 {gsub("cpu ",""); print}' /proc/stat > /tmp/.cpu2.$$
|
||||
read u1 n1 s1 i1 io1 irq1 si1 st1 g1 gn1 < /tmp/.cpu1.$$
|
||||
read u2 n2 s2 i2 io2 irq2 si2 st2 g2 gn2 < /tmp/.cpu2.$$
|
||||
rm -f /tmp/.cpu1.$$ /tmp/.cpu2.$$
|
||||
t1=$((u1+n1+s1+i1+io1+irq1+si1+st1))
|
||||
t2=$((u2+n2+s2+i2+io2+irq2+si2+st2))
|
||||
di=$((i2-i1))
|
||||
dt=$((t2-t1))
|
||||
if [ $dt -gt 0 ]; then usage=$(( (1000*(dt-di)/dt+5)/10 )); else usage=0; fi
|
||||
echo "CPU_USAGE=$usage"
|
||||
echo "CPU_CORES=$(nproc 2>/dev/null || echo 1)"
|
||||
echo "LOAD=$(cat /proc/loadavg | awk '{print $1,$2,$3}')"
|
||||
echo "UPTIME=$(awk '{printf "%.0f",$1}' /proc/uptime)"
|
||||
echo "===MEM==="
|
||||
mem_total=$(awk '/MemTotal/{print $2}' /proc/meminfo)
|
||||
mem_avail=$(awk '/MemAvailable/{print $2}' /proc/meminfo)
|
||||
swap_total=$(awk '/SwapTotal/{print $2}' /proc/meminfo)
|
||||
swap_free=$(awk '/SwapFree/{print $2}' /proc/meminfo)
|
||||
if [ -z "$mem_avail" ]; then mem_avail=$((mem_total - $(awk '/^(Buffers|Cached|SReclaimable):/{s+=$2} END{print s}' /proc/meminfo))); fi
|
||||
used=$((mem_total - mem_avail))
|
||||
echo "MEM_TOTAL=$mem_total"
|
||||
echo "MEM_USED=$used"
|
||||
echo "MEM_AVAIL=$mem_avail"
|
||||
echo "SWAP_TOTAL=$swap_total"
|
||||
echo "SWAP_USED=$((swap_total-swap_free))"
|
||||
echo "===DISK==="
|
||||
df -PB1 -x tmpfs -x devtmpfs 2>/dev/null | awk 'NR>1 {printf "DISK|%s|%d|%d|%s\n",$NF,$2,$3,$5}'
|
||||
echo "===NET==="
|
||||
for iface in $(ls /sys/class/net/ 2>/dev/null | grep -v lo); do
|
||||
rx=$(cat /sys/class/net/$iface/statistics/rx_bytes 2>/dev/null || echo 0)
|
||||
tx=$(cat /sys/class/net/$iface/statistics/tx_bytes 2>/dev/null || echo 0)
|
||||
echo "NET|$iface|$rx|$tx"
|
||||
done
|
||||
echo "===PROC==="
|
||||
ps -e -o stat= 2>/dev/null | awk '{r+=($1~/^R/); s+=($1~/^S/); d+=($1~/^D/); z+=($1~/^Z/); t++} END{printf "PROC_TOTAL=%d\nPROC_RUNNING=%d\nPROC_SLEEP=%d\nPROC_DISK=%d\nPROC_ZOMBIE=%d\n", t, r, s, d, z}'
|
||||
ps -eo pid,ppid,user,pcpu,pmem,vsz,rss,stat,pri,nice,etimes,times,args --sort=-pcpu --no-headers 2>/dev/null \
|
||||
| head -500 \
|
||||
| awk '{
|
||||
out="PROC\t";
|
||||
for(i=1;i<=12;i++) out=out $i "\t";
|
||||
rest="";
|
||||
for(i=13;i<=NF;i++) rest=(i==13?$i:rest " " $i);
|
||||
print out rest
|
||||
}'
|
||||
echo "===HOST==="
|
||||
echo "HOSTNAME=$(hostname)"
|
||||
echo "KERNEL=$(uname -r)"
|
||||
echo "OS=$(. /etc/os-release 2>/dev/null && echo "$PRETTY_NAME" || uname -s)"
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
@@ -77,7 +139,7 @@ echo "OS=$(. /etc/os-release 2>/dev/null && echo "$PRETTY_NAME" || uname -s)"
|
||||
return m.group(1).strip() if m else default
|
||||
|
||||
@classmethod
|
||||
def collect(cls, conn: SSHConnection) -> dict:
|
||||
def collect(cls, conn: SSHConnection, fast: bool = True) -> dict:
|
||||
"""采集一次指标；返回 dict"""
|
||||
empty = {
|
||||
"cpu": 0.0, "cores": 1, "load1": 0, "load5": 0, "load15": 0,
|
||||
@@ -91,7 +153,8 @@ echo "OS=$(. /etc/os-release 2>/dev/null && echo "$PRETTY_NAME" || uname -s)"
|
||||
}
|
||||
if not conn or not conn.connected:
|
||||
return empty
|
||||
code, out, err = conn.exec_command(cls._LINUX_METRICS_SCRIPT, timeout=10)
|
||||
script = cls._FAST_METRICS_SCRIPT if fast else cls._LINUX_METRICS_SCRIPT
|
||||
code, out, err = conn.exec_command(script, timeout=10)
|
||||
if code != 0 or not out:
|
||||
empty["error"] = err or "采集失败"
|
||||
return empty
|
||||
|
||||
Reference in New Issue
Block a user