#!/usr/bin/python
import collectd
import string
import sys
from NagAconda import Plugin
MyPlugin = Plugin("Plugin to check memory usage from collectd", "1.0")
MyPlugin.add_option('H', 'host', 'host to check.', required=True)
MyPlugin.add_option('S','socket','Socket to connect to. (default=/var/run/collectd-unixsock)',required=False,default='/var/run/collectd-unixsock')
MyPlugin.enable_status('warning')
MyPlugin.enable_status('critical')
MyPlugin.start()
formula = "(#"+MyPlugin.options.host+"/memory/memory-free 0# + #"+MyPlugin.options.host+"/memory/memory-cached 0# + #"+MyPlugin.options.host+"/memory/memory-buffered 0# ) / ( #"+MyPlugin.options.host+"/memory/memory-free 0# + #t430s-fpg/memory/memory-cached 0# + #"+MyPlugin.options.host+"/memory/memory-buffered 0# + #"+MyPlugin.options.host+"/memory/memory-used 0#)*100"
c = collectd.Collect(MyPlugin.options.socket)
val=c.calculate(formula)
MyPlugin.set_value('mem_free', val, scale='%')
MyPlugin.finish()
try:
content = json.loads(response.content)
except Exception as e:
check_api.unknown_error("%s health check response was malformed: %s" % (check_api.options.action, e))
else:
response = requests.get("%s:%d/api/tasks/?limit=%d" % (check_api.options.host, int(check_api.options.port), check_api.options.limit))
try:
response.raise_for_status()
except Exception as e:
print "Status Critical, task list for node %s cannot be retrieved" % check_api.options.node
sys.exit(2)
try:
content = json.loads(response.content)
except Exception as e:
check_api.unknown_error("%s health check response was malformed: %s" % (check_api.options.action, e))
failed = []
for task in content:
if task[1]["failed"]:
failed.append(task[0])
if failed:
print "Status Warning, the last %d tasks for node %s contain failures: %s" % (check_api.options.limit, check_api.options.node, failed)
sys.exit(1)
check_api.set_status_message("Celery health check successful")
check_api.finish()
response = requests.get(
"%s:%d/api/tasks?limit=%d" %
(check_api.options.host, int(
check_api.options.port), check_api.options.limit))
try:
response.raise_for_status()
except Exception as e:
print "Status Critical, task list for node %s cannot be retrieved" % check_api.options.node
sys.exit(2)
try:
content = json.loads(response.content)
except Exception as e:
check_api.unknown_error("%s health check response was malformed: %s" %
(check_api.options.action, e))
failed = []
for task in content.items():
if task[1]['failed']:
failed.append(task[0])
if failed:
print "Status Warning, the last %d tasks for node %s contain failures: %s" % (
check_api.options.limit, check_api.options.node, failed)
sys.exit(1)
check_api.set_status_message("Celery health check successful")
check_api.finish()
gmetad_port=plugin.options.gmetad_port,
gmetad_timeout=plugin.options.gmetad_timeout,
cache_path=plugin.options.cache_path,
cache_ttl=plugin.options.cache_ttl,
cache_ttl_splay=plugin.options.cache_ttl_splay,
cache_grace=plugin.options.cache_grace,
metrics_max_age=plugin.options.metrics_max_age,
debug_level=plugin.options.verbose).get_value(
metric_host=plugin.options.metric_host,
metric_name=plugin.options.metric_name)
plugin.set_status_message(
'%s = %s %s' % (value['title'], value['value'], value['units']))
if value['units'].upper() in ('B', 'KB', 'MB', 'GB', 'TB') or \
value['units'].lower() in ('s', 'ms', 'us', 'ns', '%'):
plugin.set_value(plugin.options.metric_name,
value['value'],
scale=value['units'])
else:
plugin.set_value(plugin.options.metric_name, value['value'])
except (GangliaMetrics.MetricNotFoundError), e:
plugin.unknown_error(str(e))
except (Exception), e:
print(e)
sys.exit(2)
# Print results
plugin.finish()
gmetad_port=plugin.options.gmetad_port,
gmetad_timeout=plugin.options.gmetad_timeout,
cache_path=plugin.options.cache_path,
cache_ttl=plugin.options.cache_ttl,
cache_ttl_splay=plugin.options.cache_ttl_splay,
cache_grace=plugin.options.cache_grace,
metrics_max_age=plugin.options.metrics_max_age,
debug_level=plugin.options.verbose).get_value(
metric_host=plugin.options.metric_host,
metric_name=plugin.options.metric_name)
plugin.set_status_message('%s = %s %s' % (value['title'],
value['value'],
value['units']))
if value['units'].upper() in ('B', 'KB', 'MB', 'GB', 'TB') or \
value['units'].lower() in ('s', 'ms', 'us', 'ns', '%'):
plugin.set_value(plugin.options.metric_name, value['value'],
scale=value['units'])
else:
plugin.set_value(plugin.options.metric_name, value['value'])
except (GangliaMetrics.MetricNotFoundError), e:
plugin.unknown_error(str(e))
except (Exception), e:
print(e)
sys.exit(2)
# Print results
plugin.finish()
system2_bytes_size_float = calc_bnumber(btrfs_output_parsed.system2.bytesize_total)
metadata_bytes_used_float = calc_bnumber(btrfs_output_parsed.metadata.bytesize_used)
metadata_bytes_size_float = calc_bnumber(btrfs_output_parsed.metadata.bytesize_total)
# DEBUG:
# print btrfs_output_parsed.dump()
# set nagios output
btrfs_check.set_range('warning', 100000000000000000000000, range_num=2)
btrfs_check.set_range('critical', 200000000000000000000000, range_num=2)
btrfs_check.set_value("data_used", btrfs_output_parsed.data.bytesize_used.num, scale=btrfs_output_parsed.data.bytesize_used.unit, threshold=2)
btrfs_check.set_value("data_total", btrfs_output_parsed.data.bytesize_total.num, scale=btrfs_output_parsed.data.bytesize_total.unit, threshold=2)
btrfs_check.set_value("data_ratio", data_usage_percentage_float, scale="%", threshold=1)
btrfs_check.set_value("system1_used", btrfs_output_parsed.system1.bytesize_used.num, scale=btrfs_output_parsed.system1.bytesize_used.unit, threshold=2)
btrfs_check.set_value("system1_total", btrfs_output_parsed.system1.bytesize_total.num, scale=btrfs_output_parsed.system1.bytesize_total.unit, threshold=2)
btrfs_check.set_value("system2_used", btrfs_output_parsed.system2.bytesize_used.num, scale=btrfs_output_parsed.system2.bytesize_used.unit, threshold=2)
btrfs_check.set_value("system2_total", btrfs_output_parsed.system2.bytesize_total.num, scale=btrfs_output_parsed.system2.bytesize_total.unit, threshold=2)
btrfs_check.set_value("metadata_used", btrfs_output_parsed.metadata.bytesize_used.num, scale=btrfs_output_parsed.metadata.bytesize_used.unit, threshold=2)
btrfs_check.set_value("metadata_total", btrfs_output_parsed.metadata.bytesize_total.num, scale=btrfs_output_parsed.metadata.bytesize_total.unit, threshold=2)
btrfs_check.set_status_message("{0}{1} of {2}{3} used ({4:.2%})".format(btrfs_output_parsed.data.bytesize_used.num, btrfs_output_parsed.data.bytesize_used.unit, btrfs_output_parsed.data.bytesize_total.num, btrfs_output_parsed.data.bytesize_total.unit, data_usage_ratio_float ))
btrfs_check.finish()
g.start()
# Bounds checking on crit and warn
if g.options.raw_critical < g.options.raw_warning:
g.unknown_error("ERROR: Critical level (%s) is set LOWER than Warning level (%s)" % (
g.options.raw_critical,
g.options.raw_warning,
))
# Build url
# FIXME: pickle seems efficient, but maybe harder to debug?
url = 'http://%s/render?from=%s&target=%s&format=pickle' % (
g.options.host,
g.options.window,
g.options.target,
)
if debug: print 'DEBUG URL:',url
value=get_value(url)
if debug: print 'DEBUG VALUE:', value
# Error parsing
if value == None:
g.unknown_error("ERROR: Could not parse data from URL - %s" % url)
# Set it and forget it
g.set_value(g.options.target, float(value))
g.set_status_message("%s = %f" % (g.options.target, value))
g.finish()
url = graphite.options.url
if not url:
url = ''.join([
'http://localhost/render?format=raw',
'&from=-', graphite.options.minute, 'minutes',
'&target=servers.', graphite.options.hostMafia.replace('.', '_'),
'_9400.', graphite.options.keyMafia,
])
try:
usock = urllib2.urlopen(url)
data = usock.read()
usock.close()
assert data
pieces = data.split("|")
counter = pieces[0].split(",")[0]
values = pieces[1].split(",")[:-1]
values = map(lambda x: 0.0 if x == 'None' else float(x), values)
assert not any(map(math.isnan, values))
avg = sum(values)/len(values);
graphite.set_value(counter, avg)
graphite.set_status_message("Avg value of %s was %f" % (counter, avg))
graphite.finish()
except Exception, e:
graphite.unknown_error("Error: %s" % e)