X-Git-Url: https://osm.etsi.org/gitweb/?a=blobdiff_plain;f=plugins%2FOpenStack%2FAodh%2Falarming.py;h=d409d71cb4b9103567836fbe412ccfcf49f87eb8;hb=95b92b3dacfd93fa1649a5f87dafd2fa6553a086;hp=f530a35033f0378b5cdf8beb7b7b873912414d2c;hpb=cda5f2f5df14f015829cd733d794223c0c370c41;p=osm%2FMON.git diff --git a/plugins/OpenStack/Aodh/alarming.py b/plugins/OpenStack/Aodh/alarming.py index f530a35..d409d71 100644 --- a/plugins/OpenStack/Aodh/alarming.py +++ b/plugins/OpenStack/Aodh/alarming.py @@ -22,7 +22,8 @@ """Carry out alarming requests via Aodh API.""" import json -import logging as log +import logging +log = logging.getLogger(__name__) from core.message_bus.producer import KafkaProducer @@ -33,25 +34,30 @@ from plugins.OpenStack.response import OpenStack_Response __author__ = "Helena McGough" -ALARM_NAMES = [ - "Average_Memory_Usage_Above_Threshold", - "Read_Latency_Above_Threshold", - "Write_Latency_Above_Threshold", - "DISK_READ_OPS", - "DISK_WRITE_OPS", - "DISK_READ_BYTES", - "DISK_WRITE_BYTES", - "Net_Packets_Dropped", - "Packets_in_Above_Threshold", - "Packets_out_Above_Threshold", - "CPU_Utilization_Above_Threshold"] +ALARM_NAMES = { + "average_memory_usage_above_threshold": "average_memory_utilization", + "disk_read_ops": "disk_read_ops", + "disk_write_ops": "disk_write_ops", + "disk_read_bytes": "disk_read_bytes", + "disk_write_bytes": "disk_write_bytes", + "net_packets_dropped": "packets_dropped", + "packets_in_above_threshold": "packets_received", + "packets_out_above_threshold": "packets_sent", + "cpu_utilization_above_threshold": "cpu_utilization"} SEVERITIES = { - "WARNING": "low", - "MINOR": "low", - "MAJOR": "moderate", - "CRITICAL": "critical", - "INDETERMINATE": "critical"} + "warning": "low", + "minor": "low", + "major": "moderate", + "critical": "critical", + "indeterminate": "critical"} + +STATISTICS = { + "average": "avg", + "minimum": "min", + "maximum": "max", + "count": "count", + "sum": "sum"} class Alarming(object): @@ -86,7 +92,7 @@ class Alarming(object): log.info("Alarm action required: %s" % (message.topic)) # Generate and auth_token and endpoint for request - auth_token, endpoint = self.authenticate(values) + auth_token, endpoint = self.authenticate() if message.key == "create_alarm_request": # Configure/Update an alarm @@ -97,10 +103,14 @@ class Alarming(object): # Generate a valid response message, send via producer try: + if alarm_status is True: + log.info("Alarm successfully created") + resp_message = self._response.generate_response( 'create_alarm_response', status=alarm_status, alarm_id=alarm_id, cor_id=alarm_details['correlation_id']) + log.info("Response Message: %s", resp_message) self._producer.create_alarm_response( 'create_alarm_resonse', resp_message, 'alarm_response') @@ -111,34 +121,16 @@ class Alarming(object): # Check for a specifed: alarm_name, resource_uuid, severity # and generate the appropriate list list_details = values['alarm_list_request'] - try: - name = list_details['alarm_name'] - alarm_list = self.list_alarms( - endpoint, auth_token, alarm_name=name) - except Exception as a_name: - log.debug("No name specified for list:%s", a_name) - try: - resource = list_details['resource_uuid'] - alarm_list = self.list_alarms( - endpoint, auth_token, resource_id=resource) - except Exception as r_id: - log.debug("No resource id specified for this list:\ - %s", r_id) - try: - severe = list_details['severity'] - alarm_list = self.list_alarms( - endpoint, auth_token, severity=severe) - except Exception as exc: - log.warn("No severity specified for list: %s.\ - will return full list.", exc) - alarm_list = self.list_alarms( - endpoint, auth_token) + + alarm_list = self.list_alarms( + endpoint, auth_token, list_details) try: # Generate and send a list response back resp_message = self._response.generate_response( 'list_alarm_response', alarm_list=alarm_list, cor_id=list_details['correlation_id']) + log.info("Response Message: %s", resp_message) self._producer.list_alarm_response( 'list_alarm_response', resp_message, 'alarm_response') @@ -158,6 +150,7 @@ class Alarming(object): 'delete_alarm_response', alarm_id=alarm_id, status=resp_status, cor_id=request_details['correlation_id']) + log.info("Response message: %s", resp_message) self._producer.delete_alarm_response( 'delete_alarm_response', resp_message, 'alarm_response') @@ -190,6 +183,7 @@ class Alarming(object): 'update_alarm_response', alarm_id=alarm_id, cor_id=alarm_details['correlation_id'], status=status) + log.info("Response message: %s", resp_message) self._producer.update_alarm_response( 'update_alarm_response', resp_message, 'alarm_response') @@ -209,23 +203,34 @@ class Alarming(object): url = "{}/v2/alarms/".format(endpoint) # Check if the desired alarm is supported - alarm_name = values['alarm_name'] - if alarm_name not in ALARM_NAMES: + alarm_name = values['alarm_name'].lower() + metric_name = values['metric_name'].lower() + resource_id = values['resource_uuid'] + + if alarm_name not in ALARM_NAMES.keys(): log.warn("This alarm is not supported, by a valid metric.") return None, False + if ALARM_NAMES[alarm_name] != metric_name: + log.warn("This is not the correct metric for this alarm.") + return None, False + + # Check for the required metric + metric_id = self.check_for_metric(auth_token, metric_name, resource_id) try: - metric_name = values['metric_name'] - resource_id = values['resource_uuid'] - # Check the payload for the desired alarm - payload = self.check_payload(values, metric_name, resource_id, - alarm_name) - new_alarm = self._common._perform_request( - url, auth_token, req_type="post", payload=payload) - - return json.loads(new_alarm.text)['alarm_id'], True + if metric_id is not None: + # Create the alarm if metric is available + payload = self.check_payload(values, metric_name, resource_id, + alarm_name) + new_alarm = self._common._perform_request( + url, auth_token, req_type="post", payload=payload) + return json.loads(new_alarm.text)['alarm_id'], True + else: + log.warn("The required Gnocchi metric does not exist.") + return None, False + except Exception as exc: - log.warn("Alarm creation could not be performed: %s", exc) + log.warn("Failed to create the alarm: %s", exc) return None, False def delete_alarm(self, endpoint, auth_token, alarm_id): @@ -236,6 +241,7 @@ class Alarming(object): result = self._common._perform_request( url, auth_token, req_type="delete") if str(result.status_code) == "404": + log.info("Alarm doesn't exist: %s", result.status_code) # If status code is 404 alarm did not exist return False else: @@ -245,34 +251,81 @@ class Alarming(object): log.warn("Failed to delete alarm: %s because %s.", alarm_id, exc) return False - def list_alarms(self, endpoint, auth_token, - alarm_name=None, resource_id=None, severity=None): + def list_alarms(self, endpoint, auth_token, list_details): """Generate the requested list of alarms.""" url = "{}/v2/alarms/".format(endpoint) - alarm_list = [] - - result = self._common._perform_request( - url, auth_token, req_type="get") - if result is not None: - # Check for a specified list based on: - # alarm_name, severity, resource_id - if alarm_name is not None: - for alarm in json.loads(result.text): - if alarm_name in str(alarm): - alarm_list.append(str(alarm)) - elif resource_id is not None: - for alarm in json.loads(result.text): - if resource_id in str(alarm): - alarm_list.append(str(alarm)) - elif severity is not None: + a_list, name_list, sev_list, res_list = [], [], [], [] + + # TODO(mcgoughh): for now resource_id is a mandatory field + resource = list_details['resource_uuid'] + + # Checking what fields are specified for a list request + try: + name = list_details['alarm_name'].lower() + if name not in ALARM_NAMES.keys(): + log.warn("This alarm is not supported, won't be used!") + name = None + except KeyError as exc: + log.info("Alarm name isn't specified.") + name = None + + try: + severity = list_details['severity'].lower() + sev = SEVERITIES[severity] + except KeyError as exc: + log.info("Severity is unspecified/incorrectly configured") + sev = None + + # Perform the request to get the desired list + try: + result = self._common._perform_request( + url, auth_token, req_type="get") + + if result is not None: + # Get list based on resource id for alarm in json.loads(result.text): - if severity in str(alarm): - alarm_list.append(str(alarm)) + rule = alarm['gnocchi_resources_threshold_rule'] + if resource == rule['resource_id']: + res_list.append(str(alarm)) + if not res_list: + log.info("No alarms for this resource") + return a_list + + # Generate specified listed if requested + if name is not None and sev is not None: + log.info("Return a list of %s alarms with %s severity.", + name, sev) + for alarm in json.loads(result.text): + if name == alarm['name']: + name_list.append(str(alarm)) + for alarm in json.loads(result.text): + if sev == alarm['severity']: + sev_list.append(str(alarm)) + name_sev_list = list(set(name_list).intersection(sev_list)) + a_list = list(set(name_sev_list).intersection(res_list)) + elif name is not None: + log.info("Returning a %s list of alarms.", name) + for alarm in json.loads(result.text): + if name == alarm['name']: + name_list.append(str(alarm)) + a_list = list(set(name_list).intersection(res_list)) + elif sev is not None: + log.info("Returning %s severity alarm list.", sev) + for alarm in json.loads(result.text): + if sev == alarm['severity']: + sev_list.append(str(alarm)) + a_list = list(set(sev_list).intersection(res_list)) + else: + log.info("Returning an entire list of alarms.") + a_list = res_list else: - alarm_list = result.text - else: + log.info("There are no alarms!") + + except Exception as exc: + log.info("Failed to generate required list: %s", exc) return None - return alarm_list + + return a_list def update_alarm_state(self, endpoint, auth_token, alarm_id): """Set the state of an alarm to ok when ack message is received.""" @@ -303,7 +356,7 @@ class Alarming(object): metric_name = rule['metric'] except Exception as exc: log.warn("Failed to retreive existing alarm info: %s.\ - Can only update OSM created alarms.", exc) + Can only update OSM alarms.", exc) return None, False # Generates and check payload configuration for alarm update @@ -327,12 +380,13 @@ class Alarming(object): """Check that the payload is configuration for update/create alarm.""" try: # Check state and severity - severity = values['severity'] - if severity == "INDETERMINATE": + severity = values['severity'].lower() + if severity == "indeterminate": alarm_state = "insufficient data" if alarm_state is None: alarm_state = "ok" + statistic = values['statistic'].lower() # Try to configure the payload for the update/create request # Can only update: threshold, operation, statistic and # the severity of the alarm @@ -341,7 +395,7 @@ class Alarming(object): 'metric': metric_name, 'resource_id': resource_id, 'resource_type': 'generic', - 'aggregation_method': values['statistic'].lower()} + 'aggregation_method': STATISTICS[statistic]} payload = json.dumps({'state': alarm_state, 'name': alarm_name, 'severity': SEVERITIES[severity], @@ -352,20 +406,16 @@ class Alarming(object): log.warn("Alarm is not configured correctly: %s", exc) return None - def authenticate(self, values): + def authenticate(self): """Generate an authentication token and endpoint for alarm request.""" try: # Check for a tenant_id - auth_token = self._common._authenticate( - tenant_id=values['tenant_uuid']) - endpoint = self._common.get_endpoint("alarming") - except Exception as exc: - log.warn("Tenant ID is not specified. Will use a generic\ - authentication: %s", exc) auth_token = self._common._authenticate() endpoint = self._common.get_endpoint("alarming") - - return auth_token, endpoint + return auth_token, endpoint + except Exception as exc: + log.warn("Authentication to Keystone failed:%s", exc) + return None, None def get_alarm_state(self, endpoint, auth_token, alarm_id): """Get the state of the alarm.""" @@ -378,3 +428,23 @@ class Alarming(object): except Exception as exc: log.warn("Failed to get the state of the alarm:%s", exc) return None + + def check_for_metric(self, auth_token, m_name, r_id): + """Check for the alarm metric.""" + try: + endpoint = self._common.get_endpoint("metric") + + url = "{}/v1/metric/".format(endpoint) + metric_list = self._common._perform_request( + url, auth_token, req_type="get") + + for metric in json.loads(metric_list.text): + name = metric['name'] + resource = metric['resource_id'] + if (name == m_name and resource == r_id): + metric_id = metric['id'] + log.info("The required metric exists, an alarm will be created.") + return metric_id + except Exception as exc: + log.info("Desired Gnocchi metric not found:%s", exc) + return None