def check(zone=None, slowdown=80, _max=90):
# search for hottest or use given zone
hottest = 0
for x in [zone] if zone else range(3):
try:
with open("/sys/class/thermal/thermal_zone%s/temp" % x) as f:
reading = f.read()
value = int(reading.strip()[:-3])
if value > hottest:
hottest = value
except OSError:
continue
if not hottest:
raise Exception("Could not find any thermal zone")
status = {
"value": "%s°C" % hottest,
"bar_min": "15°C",
"bar_max": "%s°C" % _max,
"bar_percent": bar_percent(hottest, _max, 15),
}
if hottest < slowdown:
status["healthy"] = True
status["reason"] = "CPU temperature nominal"
else:
status["healthy"] = False
status["reason"] = "CPU overheated"
return status
def check(_min=216, _max=253):
info = get_ups_data()
return {
"bar_min": "%sV" % _min,
"bar_max": "%sV" % _max,
"bar_percent": bar_percent(info["LINEV"], _max, _min),
"value": "%sV" % info["LINEV"],
"healthy": (info["LINEV"] < _max) and (info["LINEV"] > _min),
}
def check(host, _min=10, _max=35):
r = requests.get("http://%s/fresh.xml" % host, timeout=5)
r.raise_for_status()
root = ET.fromstring(r.text)
value = root[0].attrib["val"]
value = int(float(value))
return {
"value": "%s°C" % value,
"bar_min": "%s°C" % _min,
"bar_max": "%s°C" % _max,
"bar_percent": bar_percent(value, _max, _min),
"healthy": value < _max and value > _min,
}
def check(slowdown=88, _max=93):
result = run(
["nvidia-smi", "-q", "-d", "TEMPERATURE"],
timeout=10,
check=True,
stdout=PIPE,
)
state = dict()
for line in result.stdout.decode().splitlines():
try:
key, val = line.split(":")
except ValueError:
continue
key, val = key.strip(), val.strip()
state[key] = val
value = int(state["GPU Current Temp"][:-2])
try:
slowdown = int(state["GPU Slowdown Temp"][:-2])
_max = int(state["GPU Shutdown Temp"][:-2])
except ValueError:
# not specified, so dot change the defaults
pass
status = {
"value": "%s°C" % value,
"bar_min": "15°C",
"bar_max": "%s°C" % _max,
"bar_percent": bar_percent(value, _max, 15),
}
if value < slowdown:
status["healthy"] = True
status["reason"] = "GPU temperature nominal"
else:
status["healthy"] = False
status["reason"] = "GPU overheated"
return status
def check(_max=300):
# "return normalised % load (avg num of processes waiting per processor)"
load = os.getloadavg()[0]
load = load / multiprocessing.cpu_count()
value = int(load * 100)
status = {
"value": "%s%%" % value,
"bar_min": "0%",
"bar_max": "100%",
"bar_percent": bar_percent(value, 100),
"healthy": value < _max,
}
if value < _max:
status["healthy"] = True
status["reason"] = "CPU usage nominal"
else:
status["healthy"] = False
status["reason"] = "CPU overloaded"
return status
def check(mountpoint="/"):
s = os.statvfs(mountpoint)
free = s.f_bsize * s.f_bavail
total = s.f_bsize * s.f_blocks
usage = total - free
percent = bar_percent(usage, total)
status = {
"value": human_bytes(usage),
"bytes": usage,
"bar_min": "0 GB",
"bar_max": human_bytes(total),
"bar_percent": percent,
}
if usage < 0.9 * total:
status["healthy"] = True
status["reason"] = "Disk usage nominal"
else:
status["healthy"] = False
status["reason"] = "Disk is nearly full"
return status
def check():
# http://www.linuxatemyram.com/
with open("/proc/meminfo") as f:
lines = f.readlines()
# in kB
info = {}
for line in lines:
m = re.search("(\w+):\s*(\d+)", line)
if m:
info[m.group(1)] = int(m.group(2))
used = info["MemTotal"] - info["MemFree"] - info["Buffers"] - info["Cached"]
total = info["MemTotal"] * 1024
# used by applications, not cache/buffers
value = used * 1024
status = {
"value": human_bytes(value),
"bytes": value,
"bar_min": "0 GB",
"bar_max": human_bytes(total),
"bar_percent": bar_percent(value, total),
}
if value < 0.9 * total:
status["healthy"] = True
status["reason"] = "RAM usage nominal"
else:
status["healthy"] = False
status["reason"] = "RAM usage too high"
return status