def main():
'''get docker and openshift versions and send to metric sender
'''
args = parse_args()
mts = MetricSender(verbose=args.verbose, debug=args.debug)
# Check if host rpm db is mounted. Otherwise check againts container db
rpm_db_path = "/host/var/lib/rpm"
if not os.path.exists(rpm_db_path):
rpm_db_path = "/var/lib/rpm"
keys = {}
# Accumulate failures
failures = 0
# Get docker version
success, err = add_specific_rpm_version("docker", rpm_db_path, keys, mts)
if not success:
failures += 1
print "Failed to get docker rpm version. " + err.output
openshift_package_name = "origin"
# Get openshift node version (attempt upstream)
success, err = add_specific_rpm_version(
"{}-node".format(openshift_package_name), rpm_db_path, keys, mts,
"openshift.node.")
if not success:
# Get openshift version (attempt downstream)
openshift_package_name = "atomic-openshift"
success, err2 = add_specific_rpm_version(
"{}-node".format(openshift_package_name), rpm_db_path, keys, mts,
"openshift.node.")
if not success:
failures += 1
print "Failed to get openshift rpm version:\n" + err.output + +err2.output
# Get openshift master version (upstream or downstream) - only if node rpm found
if success:
success, err = add_specific_rpm_version(
"{}-master".format(openshift_package_name), rpm_db_path, keys, mts,
"openshift.master.")
if not success:
# Print notification but don't count this as failure
print "Note: " + err.output
print "Sending these metrics:"
print json.dumps(keys, indent=4)
mts.send_metrics()
print "\nDone.\n"
sys.exit(failures)
def send_metrics(usage, capacity, used):
""" send data to MetricSender"""
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
ms.add_metric({'openshift.master.pv.percent.usage': usage})
ms.add_metric({'openshift.master.pv.capacity.max': capacity})
ms.add_metric({'openshift.master.pv.capacity.used': used})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def main():
""" ds pod check """
args = parse_args()
logger.debug("args: ")
logger.debug(args)
result = check_ds_status(args=args, )
#send the value to zabbix
mts = MetricSender(verbose=args.verbose)
mts.add_metric({args.key: result})
mts.send_metrics()
def main():
""" SAML Pod Status """
args = parse_args()
logger.debug("args: ")
logger.debug(args)
result = test_saml_pod(args=args, )
#send the value to zabbix
mts = MetricSender(verbose=args.verbose)
mts.add_metric({'openshift.master.saml.status': result})
mts.send_metrics()
def run(self):
""" Main function to run the check """
self.parse_args()
self.metric_sender = MetricSender(verbose=self.args.verbose,
debug=self.args.debug)
status = self.parse_config()
self.metric_sender.add_metric({"openshift.kubeconfig.status": status})
self.metric_sender.send_metrics()
def main():
""" check all the node pods tatus see if any node have problem """
args = parse_args()
logger.debug("args: ")
logger.debug(args)
nsList = args.skip_namespaces
pods_status = check_node_pods_status(nsList)
#send the value to zabbix
mts = MetricSender(verbose=args.verbose)
mts.add_metric({'openshift.nodes.pods.status': pods_status})
mts.send_metrics()
def report_to_zabbix(total_expired_snapshots, total_deleted_snapshots,
total_deletion_errors):
""" Sends the commands exit code to zabbix. """
mts = MetricSender(verbose=True)
mts.add_metric({
EXPIRED_SNAPSHOTS_KEY: total_expired_snapshots,
DELETED_SNAPSHOTS_KEY: total_deleted_snapshots,
DELETION_ERRORS_KEY: total_deletion_errors
})
mts.send_metrics()
def main():
""" check all the node taints see if any node have problem """
args = parse_args()
logger.debug("args: ")
logger.debug(args)
taint_status = check_taint_status()
#send the value to zabbix
mts = MetricSender(verbose=args.verbose)
mts.add_metric({'openshift.nodes.taint.status': taint_status})
mts.send_metrics()
def main():
""" Main function to run the check """
args = parse_args()
metric_sender = MetricSender(verbose=args.verbose, debug=args.debug)
discovery_key_disk = 'disc.disk'
interval = 3
pcp_disk_dev_metrics = ['disk.dev.total', 'disk.dev.avactive']
item_prototype_macro_disk = '#OSO_DISK'
item_prototype_key_tps = 'disc.disk.tps'
item_prototype_key_putil = 'disc.disk.putil'
disk_metrics = pminfo.get_sampled_data(pcp_disk_dev_metrics, interval, 2)
pcp_metrics_divided = {}
for metric in pcp_disk_dev_metrics:
pcp_metrics_divided[metric] = {
k: v
for k, v in disk_metrics.items() if metric in k
}
# do TPS checks; use disk.dev.total
filtered_disk_totals = clean_up_metric_dict(
pcp_metrics_divided[pcp_disk_dev_metrics[0]],
pcp_disk_dev_metrics[0] + '.')
# Add dynamic items
metric_sender.add_dynamic_metric(discovery_key_disk,
item_prototype_macro_disk,
filtered_disk_totals.keys())
# calculate the TPS and add them to the ZaggSender
for disk, totals in filtered_disk_totals.iteritems():
disk_tps = (totals[1] - totals[0]) / interval
metric_sender.add_metric(
{'%s[%s]' % (item_prototype_key_tps, disk): disk_tps})
# do % Util checks; use disk.dev.avactive
filtered_disk_totals = clean_up_metric_dict(
pcp_metrics_divided[pcp_disk_dev_metrics[1]],
pcp_disk_dev_metrics[1] + '.')
# calculate the % Util and add them to the ZaggSender
for disk, totals in filtered_disk_totals.iteritems():
total_active = (float)(totals[1] - totals[0]) / 1000.0
putil = 100 * total_active / interval
metric_sender.add_metric(
{'%s[%s]' % (item_prototype_key_putil, disk): putil})
metric_sender.send_metrics()
def send_metrics(reload_time):
""" send data to MetricSender"""
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
ms.add_metric({'openshift.haproxy.reload_time' : reload_time})
logger.debug({'openshift.haproxy.reload_time' : reload_time})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def main():
""" check all the node labels see if anything miss """
args = parse_args()
logger.debug("args: ")
logger.debug(args)
#result = test_saml_pod(args=args, )
label_status = check_label_status()
#send the value to zabbix
mts = MetricSender(verbose=args.verbose)
mts.add_metric({'openshift.nodes.label.status': label_status})
mts.send_metrics()
def send_metrics(key, result):
''' send data to MetricSender '''
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
ms.add_metric({key: result})
logger.debug({key: result})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def __init__(self):
""" initialize EBSStuckVolumesCheck class """
self.args = None
self.vol_state_data = None
self.parse_args()
# Make sure we're using the profile they've requested.
if self.args.aws_creds_profile:
os.environ['AWS_PROFILE'] = self.args.aws_creds_profile
self.eu = EbsUtil(self.args.region, verbose=self.args.verbose)
self.mts = MetricSender(verbose=self.args.verbose)
def run(self):
""" Main function that runs the check """
self.parse_args()
self.metric_sender = MetricSender(verbose=self.args.verbose,
debug=self.args.debug)
self.oc = OCUtil(namespace=self.get_logging_namespace(),
config_file='/tmp/admin.kubeconfig',
verbose=self.args.verbose)
self.get_pods()
oldest_buffer = self.check_fluentd_queues()
self.send_metrics(oldest_buffer)
def send_metrics(stuck_builds, build_state):
""" send data to MetricSender"""
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
logger.debug({'openshift.stuck_builds.%s' % build_state.lower() : stuck_builds})
ms.add_metric({'openshift.stuck_builds.%s' % build_state.lower() : stuck_builds})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def send_zagg_data(keep_time):
''' send data to Zagg'''
logger.debug('send_zagg_data()')
zgs_time = time.time()
zgs = MetricSender()
zgs.add_metric({'openshift.master.project.terminating.time': keep_time})
try:
zgs.send_metrics()
except:
logger.exception('Error sending to Zagg')
logger.info("Data sent in %s seconds", str(time.time() - zgs_time))
def kill(self):
''' class entrypoint '''
self.parse_args()
self.current_time = time.time()
haproxy_procs_etimes = self.get_all_haproxy_procs()
# identify most recent haproxy process
# and remove it from list of haproxy processes
try:
youngest_etimes = min(haproxy_procs_etimes.keys())
youngest_pid = haproxy_procs_etimes[youngest_etimes]
self.dprint("Youngest haproxy PID: {}".format(youngest_pid))
haproxy_procs_etimes.pop(youngest_etimes)
except ValueError:
pass
# find processes that have connections only in 'CLOSE-WAIT' state
kill_list = []
for proc in haproxy_procs_etimes.values():
try:
only_close_wait = True
process = psutil.Process(proc)
for conn in process.connections():
if conn.status != 'CLOSE_WAIT' and conn.status != 'FIN_WAIT2':
only_close_wait = False
break
if only_close_wait:
self.dprint("PID: {} marked for removal".format(proc))
kill_list.append(proc)
except psutil.NoSuchProcess:
pass
# stop processes on the kill_list
kill_count = 0
for proc in kill_list:
try:
process = psutil.Process(proc)
self.dprint("Stopping PID: {}".format(process.pid))
process.kill()
kill_count += 1
except psutil.NoSuchProcess:
pass
print "Stopped {} haproxy processes".format(kill_count)
ms = MetricSender()
ms.add_metric({ZABBIX_KEY: kill_count})
ms.send_metrics()
def send_metrics(problems):
""" send data to MetricSender"""
logger.debug("send_metrics(problems)")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
ms.add_metric({'aws.ec2.instance.instance_status': problems['InstanceStatus']})
ms.add_metric({'aws.ec2.instance.system_status': problems['SystemStatus']})
ms.add_metric({'aws.ec2.instance.events': problems['Events']})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def send_metrics(builds):
""" send data to MetricSender"""
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
for build_state, count in builds.items():
logger.debug({'openshift.build_state.%s' % build_state: count})
ms.add_metric({'openshift.build_state.%s' % build_state: count})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def send_metrics(status_code_create, status_code_delete):
""" send data to MetricSender"""
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Send data to MetricSender")
# 1 means create and delete the project failed
ms.add_metric({'openshift.master.project.create': status_code_create})
ms.add_metric({'openshift.master.project.delete': status_code_delete})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds", str(time.time() - ms_time))
def __init__(self, config_file=None):
if not config_file:
self.config_file = '/etc/openshift_tools/container_metrics.yml'
else:
self.config_file = config_file
self.config = None
self.parse_config()
self.cli = AutoVersionClient(base_url='unix://var/run/docker.sock',
timeout=120)
self.docker_util = DockerUtil(self.cli)
self.metric_sender = MetricSender(verbose=True)
def run(self):
"""Main function to run the check"""
self.ocutil = OCUtil(config_file=self.kubeconfig,
verbose=self.args.verbose)
self.metrics = MetricSender(verbose=self.args.verbose,
debug=self.args.debug)
self.check_all_router_health()
self.check_router_replica_count()
if self.args.dry_run:
self.metrics.print_unique_metrics_key_value()
else:
self.metrics.send_metrics()
def run(self):
''' Main function that runs the check '''
self.parse_args()
self.metric_sender = MetricSender(verbose=self.args.verbose,
debug=self.args.debug)
self.oc = OCUtil(namespace='openshift-infra',
config_file=self.kubeconfig,
verbose=self.args.verbose)
pod_report = self.check_pods()
self.get_hawkular_creds()
metrics_report = self.check_node_metrics()
self.report_to_zabbix(pod_report, metrics_report)
def config_metric_sender(self):
""" configure the metric_sender """
metric_verbose = self.args.verbose
metric_debug = self.args.debug
host = self.args.host if self.args.host else self.config['host']['name']
if isinstance(metric_verbose, str):
metric_verbose = (metric_verbose == 'True')
if isinstance(metric_debug, str):
metric_debug = (metric_debug == 'True')
self.metric_sender = MetricSender(host=host, verbose=metric_verbose, debug=metric_debug,
config_file=self.args.config_file)
def run(self):
""" Main function to run the check """
self.parse_args()
self.ora = OpenshiftRestApi()
self.metric_sender = MetricSender(verbose=self.args.verbose,
debug=self.args.debug)
try:
self.get_pods()
except Exception as ex:
print "Problem retreiving pod data: %s " % ex.message
self.metric_sender.send_metrics()
def send_metrics(self, oldest_buffer):
""" Send data to MetricSender """
logger.debug("send_metrics()")
ms_time = time.time()
ms = MetricSender()
logger.info("Sending data to MetricSender...")
logger.debug({'openshift.logging.fluentd.queue.oldest': oldest_buffer})
ms.add_metric(
{'openshift.logging.fluentd.queue.oldest': oldest_buffer})
ms.send_metrics()
logger.info("Data sent to Zagg in %s seconds",
str(time.time() - ms_time))
def main():
""" Main function to run the check """
args = parse_args()
metric_sender = MetricSender(verbose=args.verbose, debug=args.debug)
filesys_full_metric = ['filesys.full']
filesys_inode_derived_metrics = {'filesys.inodes.pused' :
'filesys.usedfiles / (filesys.usedfiles + filesys.freefiles) * 100'
}
discovery_key_fs = 'disc.filesys'
item_prototype_macro_fs = '#OSO_FILESYS'
item_prototype_key_full = 'disc.filesys.full'
item_prototype_key_inode = 'disc.filesys.inodes.pused'
# Get the disk space
filesys_full_metrics = pminfo.get_metrics(filesys_full_metric)
filtered_filesys_metrics = filter_out_key_name_chars(filesys_full_metrics, 'filesys.full.')
filtered_filesys_metrics = filter_out_container_root(filtered_filesys_metrics)
if args.filter_pod_pv:
filtered_filesys_metrics = filter_out_customer_pv_filesystems(filtered_filesys_metrics)
if args.force_send_zeros:
filtered_filesys_metrics = zero_mount_percentages(filtered_filesys_metrics)
metric_sender.add_dynamic_metric(discovery_key_fs, item_prototype_macro_fs, filtered_filesys_metrics.keys())
for filesys_name, filesys_full in filtered_filesys_metrics.iteritems():
metric_sender.add_metric({'%s[%s]' % (item_prototype_key_full, filesys_name): filesys_full})
# Get filesytem inode metrics
filesys_inode_metrics = pminfo.get_metrics(derived_metrics=filesys_inode_derived_metrics)
filtered_filesys_inode_metrics = filter_out_key_name_chars(filesys_inode_metrics, 'filesys.inodes.pused.')
filtered_filesys_inode_metrics = filter_out_container_root(filtered_filesys_inode_metrics)
if args.filter_pod_pv:
filtered_filesys_inode_metrics = filter_out_customer_pv_filesystems(filtered_filesys_inode_metrics)
if args.force_send_zeros:
filtered_filesys_inode_metrics = zero_mount_percentages(filtered_filesys_inode_metrics)
for filesys_name, filesys_inodes in filtered_filesys_inode_metrics.iteritems():
metric_sender.add_metric({'%s[%s]' % (item_prototype_key_inode, filesys_name): filesys_inodes})
metric_sender.send_metrics()
def report_to_zabbix(self, disc_key, disc_macro, item_proto_key, value):
""" Sends the commands exit code to zabbix. """
mts = MetricSender()
# Add the dynamic item
self.verbose_print("Adding the dynamic item to Zabbix - %s, %s, [%s]" % \
(disc_key, disc_macro, self.args.name))
mts.add_dynamic_metric(disc_key, disc_macro, [self.args.name])
# Send the value for the dynamic item
self.verbose_print("Sending metric to Zabbix - %s[%s]: %s" % \
(item_proto_key, self.args.name, value))
mts.add_metric({'%s[%s]' % (item_proto_key, self.args.name): value})
# Actually send them
mts.send_metrics()
def send_metric_data(bucket_list, bucket_stats, args):
'''send data to zabbix '''
discovery_key = "disc.aws"
discovery_macro = "#S3_BUCKET"
prototype_s3_size = "disc.aws.size"
prototype_s3_count = "disc.aws.objects"
mts = MetricSender(verbose=args.debug)
mts.add_dynamic_metric(discovery_key, discovery_macro, bucket_list)
for bucket in bucket_stats.keys():
zab_key = "{}[{}]".format(prototype_s3_size, bucket)
mts.add_metric({zab_key: int(round(bucket_stats[bucket]["size"]))})
zab_key = "{}[{}]".format(prototype_s3_count, bucket)
mts.add_metric({zab_key: bucket_stats[bucket]["objects"]})
mts.send_metrics()
def main():
''' Gather and examine details about this node within ELBs '''
args = parse_args()
aws_access, aws_secret = get_aws_creds('/root/.aws/credentials')
instance_region = get_instance_region()
# Create boto client to access ELB resources
client = boto3.client(
'elb',
aws_access_key_id=aws_access,
aws_secret_access_key=aws_secret,
region_name=instance_region
)
# Call all available loadbalancers in the AWS account and store blob result in elb_descriptions
elb_descriptions = client.describe_load_balancers()
elb_names = map(get_elb_name, elb_descriptions['LoadBalancerDescriptions'])
# Get a list of available ELBs for a cluster
elb_tags = client.describe_tags(LoadBalancerNames=elb_names)
cluster_elbs = filter_by_cluster(elb_tags, args.clusterid)
# Filter any ELBs created by service of type LoadBalancer that is not in our watched namespaces
monitored_elbs = filter_monitored_service_elbs(cluster_elbs)
monitored_elb_names = map(get_elb_name, monitored_elbs)
# Perform health check of each instance available behind each ELB
elb_health_check(client, monitored_elb_names)
### Metric Checks
if len(elb_no_instances) != 0:
for _, elb in enumerate(elb_no_instances):
elb_instances_unhealthy.append(elb)
print "ELB: %s has no instances behind it. Please investigate." % elb
### Unhealthy count check
elb_instances_unhealthy_metric = len(elb_instances_unhealthy)
if elb_instances_unhealthy_metric != 0:
for _, unhealthy in enumerate(elb_instances_unhealthy):
print unhealthy
# ''' Now that we know if this instance is missing, feed zabbix '''
mts = MetricSender(verbose=args.verbose, debug=args.debug)
mts.add_metric({'openshift.aws.elb.health' : elb_instances_unhealthy_metric})
mts.send_metrics()