def check_thresholds(percent):
"""Gets the current state of the mesos cluster and compares it
to a given percentage. If either the ram or CPU utilization is over that
percentage, the sensu event will be sent with a status code of 2."""
stats = get_mesos_stats()
over_threshold = False
output = ""
current_mem = stats['master/mem_percent']
current_cpu = stats['master/cpus_percent']
percent = int(percent)
cpu_print_tuple = (percent, current_cpu)
mem_print_tuple = (percent, current_mem)
if current_mem >= percent:
output += "CRITICAL: Memory usage is over %d%%! Currently at %f%%!\n" % mem_print_tuple
over_threshold = True
else:
output += "OK: Memory usage is under %d%%. (Currently at %f%%)\n" % mem_print_tuple
if current_cpu >= percent:
output += "CRITICAL: CPU usage is over %d%%! Currently at %f%%!\n" % cpu_print_tuple
over_threshold = True
else:
output += "OK: CPU usage is under %d%%. (Currently at %f%%)\n" % cpu_print_tuple
if over_threshold is True:
status = 2
else:
status = 0
send_event(status, output)
return output
def check_thresholds(percent):
"""Gets the current state of the mesos cluster and compares it
to a given percentage. If either the ram or CPU utilization is over that
percentage, the sensu event will be sent with a status code of 2."""
stats = get_mesos_stats()
over_threshold = False
output = ""
current_mem = stats['master/mem_percent']
current_cpu = stats['master/cpus_percent']
percent = int(percent)
cpu_print_tuple = (percent, current_cpu)
mem_print_tuple = (percent, current_mem)
if current_mem >= percent:
output += "CRITICAL: Memory usage is over %d%%! Currently at %f%%!\n" % mem_print_tuple
over_threshold = True
else:
output += "OK: Memory usage is under %d%%. (Currently at %f%%)\n" % mem_print_tuple
if current_cpu >= percent:
output += "CRITICAL: CPU usage is over %d%%! Currently at %f%%!\n" % cpu_print_tuple
over_threshold = True
else:
output += "OK: CPU usage is under %d%%. (Currently at %f%%)\n" % cpu_print_tuple
if over_threshold is True:
status = 2
else:
status = 0
send_event(status, output)
return output
def get_mesos_status(mesos_state, verbosity):
"""Gathers information about the mesos cluster.
:return: tuple of a string containing the status and a bool representing if it is ok or not
"""
cluster_results = run_healthchecks_with_param(
mesos_state, [assert_quorum_size, assert_no_duplicate_frameworks])
metrics = get_mesos_stats()
metrics_results = run_healthchecks_with_param(metrics, [
assert_cpu_health,
assert_memory_health,
assert_disk_health,
assert_tasks_running,
assert_slave_health,
])
if verbosity == 2:
metrics_results.extend(
run_healthchecks_with_param(mesos_state,
[assert_extra_attribute_data]))
elif verbosity >= 3:
metrics_results.extend(
run_healthchecks_with_param(mesos_state,
[assert_extra_slave_data]))
return cluster_results + metrics_results
def get_mesos_status():
"""Gathers information about the mesos cluster.
:return: tuple of a string containing the status and a bool representing if it is ok or not
"""
state = get_mesos_state_from_leader()
cluster_results = run_healthchecks_with_param(state, [assert_quorum_size, assert_no_duplicate_frameworks])
metrics = get_mesos_stats()
metrics_results = run_healthchecks_with_param(metrics, [
assert_cpu_health,
assert_memory_health,
assert_slave_health,
assert_tasks_running])
return cluster_results + metrics_results
def get_mesos_status():
"""Gathers information about the mesos cluster.
:return: tuple of a string containing the status and a bool representing if it is ok or not
"""
state = get_mesos_state_from_leader()
cluster_results = run_healthchecks_with_param(
state, [assert_quorum_size, assert_no_duplicate_frameworks])
metrics = get_mesos_stats()
metrics_results = run_healthchecks_with_param(metrics, [
assert_cpu_health, assert_memory_health, assert_slave_health,
assert_tasks_running
])
return cluster_results + metrics_results
def get_mesos_status(mesos_state, verbosity):
"""Gathers information about the mesos cluster.
:return: tuple of a string containing the status and a bool representing if it is ok or not
"""
cluster_results = run_healthchecks_with_param(mesos_state, [assert_quorum_size, assert_no_duplicate_frameworks])
metrics = get_mesos_stats()
metrics_results = run_healthchecks_with_param(metrics, [
assert_cpu_health,
assert_memory_health,
assert_tasks_running,
assert_slave_health,
])
if verbosity >= 2:
metrics_results.extend(run_healthchecks_with_param(mesos_state, [assert_extra_slave_data]))
return cluster_results + metrics_results
def main():
marathon_config = None
chronos_config = None
args = parse_args()
try:
mesos_state = get_mesos_state_from_leader()
except MasterNotAvailableException as e:
# if we can't connect to master at all,
# then bomb out early
print(PaastaColors.red("CRITICAL: %s" % e.message))
sys.exit(2)
mesos_state_status = get_mesos_state_status(
mesos_state=mesos_state,
)
metrics = get_mesos_stats()
mesos_metrics_status = get_mesos_metrics_health(mesos_metrics=metrics)
all_mesos_results = mesos_state_status + mesos_metrics_status
# Check to see if Marathon should be running here by checking for config
marathon_config = marathon_tools.load_marathon_config()
# Check to see if Chronos should be running here by checking for config
chronos_config = load_chronos_config()
if marathon_config:
marathon_client = get_marathon_client(marathon_config)
try:
marathon_results = get_marathon_status(marathon_client)
except MarathonError as e:
print(PaastaColors.red("CRITICAL: Unable to contact Marathon! Error: %s" % e))
sys.exit(2)
else:
marathon_results = [HealthCheckResult(message='Marathon is not configured to run here', healthy=True)]
if chronos_config:
chronos_client = get_chronos_client(chronos_config)
try:
chronos_results = get_chronos_status(chronos_client)
except (ServerNotFoundError, socket_error) as e:
print(PaastaColors.red("CRITICAL: Unable to contact Chronos! Error: %s" % e))
sys.exit(2)
else:
chronos_results = [HealthCheckResult(message='Chronos is not configured to run here', healthy=True)]
mesos_ok = all(status_for_results(all_mesos_results))
marathon_ok = all(status_for_results(marathon_results))
chronos_ok = all(status_for_results(chronos_results))
mesos_summary = generate_summary_for_check("Mesos", mesos_ok)
marathon_summary = generate_summary_for_check("Marathon", marathon_ok)
chronos_summary = generate_summary_for_check("Chronos", chronos_ok)
healthy_exit = True if all([mesos_ok, marathon_ok, chronos_ok]) else False
if args.verbose == 0:
print mesos_summary
print marathon_summary
print chronos_summary
elif args.verbose == 1:
print mesos_summary
print_results_for_healthchecks(mesos_ok, all_mesos_results, args.verbose)
print marathon_summary
print_results_for_healthchecks(marathon_ok, marathon_results, args.verbose)
print chronos_summary
print_results_for_healthchecks(chronos_ok, chronos_results, args.verbose)
else:
print mesos_summary
print_results_for_healthchecks(mesos_ok, all_mesos_results, args.verbose)
for grouping in args.groupings:
print_with_indent('Resources Grouped by %s' % grouping, 2)
resource_info_dict = get_resource_utilization_by_grouping(key_func_for_attribute(grouping), mesos_state)
all_rows = [[grouping.capitalize(), 'CPU (free/total)', 'RAM (free/total)', 'Disk (free/total)']]
table_rows = []
for attribute_value, resource_info_dict in resource_info_dict.items():
resource_utilizations = resource_utillizations_from_resource_info(
total=resource_info_dict['total'],
free=resource_info_dict['free'],
)
healthcheck_utilization_pairs = [
healthcheck_result_resource_utilization_pair_for_resource_utilization(utilization, args.threshold)
for utilization in resource_utilizations
]
healthy_exit = all(pair[0].healthy for pair in healthcheck_utilization_pairs)
table_rows.append(get_table_rows_for_resource_info_dict(
attribute_value,
healthcheck_utilization_pairs,
args.humanize
))
table_rows = sorted(table_rows, key=lambda x: x[0])
all_rows.extend(table_rows)
for line in format_table(all_rows):
print_with_indent(line, 4)
if args.verbose == 3:
print_with_indent('Per Slave Utilization', 2)
slave_resource_dict = get_resource_utilization_by_grouping(lambda slave: slave['hostname'], mesos_state)
all_rows = [['Hostname', 'CPU (free/total)', 'RAM (free/total)', 'Disk (free/total)']]
# print info about slaves here. Note that we don't make modifications to
# the healthy_exit variable here, because we don't care about a single slave
# having high usage.
for attribute_value, resource_info_dict in slave_resource_dict.items():
table_rows = []
resource_utilizations = resource_utillizations_from_resource_info(
total=resource_info_dict['total'],
free=resource_info_dict['free'],
)
healthcheck_utilization_pairs = [
healthcheck_result_resource_utilization_pair_for_resource_utilization(utilization, args.threshold)
for utilization in resource_utilizations
]
table_rows.append(get_table_rows_for_resource_info_dict(
attribute_value,
healthcheck_utilization_pairs,
args.humanize
))
table_rows = sorted(table_rows, key=lambda x: x[0])
all_rows.extend(table_rows)
for line in format_table(all_rows):
print_with_indent(line, 4)
print marathon_summary
print_results_for_healthchecks(marathon_ok, marathon_results, args.verbose)
print chronos_summary
print_results_for_healthchecks(chronos_ok, chronos_results, args.verbose)
print "Master paasta_tools version: {0}".format(__version__)
if not healthy_exit:
sys.exit(2)
else:
sys.exit(0)
def main():
marathon_config = None
chronos_config = None
args = parse_args()
try:
mesos_state = get_mesos_state_from_leader()
except MasterNotAvailableException as e:
# if we can't connect to master at all,
# then bomb out early
print(PaastaColors.red("CRITICAL: %s" % e.message))
sys.exit(2)
mesos_state_status = get_mesos_state_status(
mesos_state=mesos_state,
)
metrics = get_mesos_stats()
mesos_metrics_status = get_mesos_metrics_health(mesos_metrics=metrics)
all_mesos_results = mesos_state_status + mesos_metrics_status
# Check to see if Marathon should be running here by checking for config
marathon_config = marathon_tools.load_marathon_config()
# Check to see if Chronos should be running here by checking for config
chronos_config = load_chronos_config()
if marathon_config:
marathon_client = get_marathon_client(marathon_config)
try:
marathon_results = get_marathon_status(marathon_client)
except MarathonError as e:
print(PaastaColors.red("CRITICAL: Unable to contact Marathon! Error: %s" % e))
sys.exit(2)
else:
marathon_results = [HealthCheckResult(message='Marathon is not configured to run here', healthy=True)]
if chronos_config:
chronos_client = get_chronos_client(chronos_config)
try:
chronos_results = get_chronos_status(chronos_client)
except (ServerNotFoundError, socket_error) as e:
print(PaastaColors.red("CRITICAL: Unable to contact Chronos! Error: %s" % e))
sys.exit(2)
else:
chronos_results = [HealthCheckResult(message='Chronos is not configured to run here', healthy=True)]
mesos_ok = all(status_for_results(all_mesos_results))
marathon_ok = all(status_for_results(marathon_results))
chronos_ok = all(status_for_results(chronos_results))
mesos_summary = generate_summary_for_check("Mesos", mesos_ok)
marathon_summary = generate_summary_for_check("Marathon", marathon_ok)
chronos_summary = generate_summary_for_check("Chronos", chronos_ok)
healthy_exit = True if all([mesos_ok, marathon_ok, chronos_ok]) else False
if args.verbose == 0:
print mesos_summary
print marathon_summary
print chronos_summary
elif args.verbose == 1:
print mesos_summary
print_results_for_healthchecks(mesos_ok, all_mesos_results, args.verbose)
print marathon_summary
print_results_for_healthchecks(marathon_ok, marathon_results, args.verbose)
print chronos_summary
print_results_for_healthchecks(chronos_ok, chronos_results, args.verbose)
elif args.verbose == 2:
print mesos_summary
print_results_for_healthchecks(mesos_ok, all_mesos_results, args.verbose)
for grouping in args.groupings:
print_with_indent('Resources Grouped by %s' % grouping, 2)
resource_info_dict = get_resource_utilization_by_grouping(key_func_for_attribute(grouping), mesos_state)
all_rows = [[grouping.capitalize(), 'CPU (free/total)', 'RAM (free/total)', 'Disk (free/total)']]
table_rows = []
for attribute_value, resource_info_dict in resource_info_dict.items():
resource_utilizations = resource_utillizations_from_resource_info(
total=resource_info_dict['total'],
free=resource_info_dict['free'],
)
healthcheck_utilization_pairs = [
healthcheck_result_resource_utilization_pair_for_resource_utilization(utilization, args.threshold)
for utilization in resource_utilizations
]
healthy_exit = all(pair[0].healthy for pair in healthcheck_utilization_pairs)
table_rows.append(get_table_rows_for_resource_info_dict(
attribute_value,
healthcheck_utilization_pairs,
args.humanize
))
table_rows = sorted(table_rows, key=lambda x: x[0])
all_rows.extend(table_rows)
for line in format_table(all_rows):
print_with_indent(line, 4)
print marathon_summary
print_results_for_healthchecks(marathon_ok, marathon_results, args.verbose)
print chronos_summary
print_results_for_healthchecks(chronos_ok, chronos_results, args.verbose)
else:
print mesos_summary
print_results_for_healthchecks(mesos_ok, all_mesos_results, args.verbose)
for grouping in args.groupings:
print_with_indent('Resources Grouped by %s' % grouping, 2)
resource_info_dict = get_resource_utilization_by_grouping(key_func_for_attribute(grouping), mesos_state)
all_rows = [[grouping.capitalize(), 'CPU (free/total)', 'RAM (free/total)', 'Disk (free/total)']]
table_rows = []
for attribute_value, resource_info_dict in resource_info_dict.items():
resource_utilizations = resource_utillizations_from_resource_info(
total=resource_info_dict['total'],
free=resource_info_dict['free'],
)
healthcheck_utilization_pairs = [
healthcheck_result_resource_utilization_pair_for_resource_utilization(utilization, args.threshold)
for utilization in resource_utilizations
]
healthy_exit = all(pair[0].healthy for pair in healthcheck_utilization_pairs)
table_rows.append(get_table_rows_for_resource_info_dict(
attribute_value,
healthcheck_utilization_pairs,
args.humanize
))
table_rows = sorted(table_rows, key=lambda x: x[0])
all_rows.extend(table_rows)
for line in format_table(all_rows):
print_with_indent(line, 4)
print_with_indent('Per Slave Utilization', 2)
slave_resource_dict = get_resource_utilization_by_grouping(lambda slave: slave['hostname'], mesos_state)
all_rows = [['Hostname', 'CPU (free/total)', 'RAM (free/total)', 'Disk (free/total)']]
# print info about slaves here. Note that we don't make modifications to
# the healthy_exit variable here, because we don't care about a single slave
# having high usage.
for attribute_value, resource_info_dict in slave_resource_dict.items():
table_rows = []
resource_utilizations = resource_utillizations_from_resource_info(
total=resource_info_dict['total'],
free=resource_info_dict['free'],
)
healthcheck_utilization_pairs = [
healthcheck_result_resource_utilization_pair_for_resource_utilization(utilization, args.threshold)
for utilization in resource_utilizations
]
table_rows.append(get_table_rows_for_resource_info_dict(
attribute_value,
healthcheck_utilization_pairs,
args.humanize
))
table_rows = sorted(table_rows, key=lambda x: x[0])
all_rows.extend(table_rows)
for line in format_table(all_rows):
print_with_indent(line, 4)
if not healthy_exit:
sys.exit(2)
else:
sys.exit(0)