3c6d4d12c436129b8ec33164831f31c034b4e637
1 # Copyright 2017 Intel Research and Development Ireland Limited
2 # *************************************************************
4 # This file is part of OSM Monitoring module
5 # All Rights Reserved to Intel Corporation
7 # Licensed under the Apache License, Version 2.0 (the "License"); you may
8 # not use this file except in compliance with the License. You may obtain
9 # a copy of the License at
11 # http://www.apache.org/licenses/LICENSE-2.0
13 # Unless required by applicable law or agreed to in writing, software
14 # distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
15 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
16 # License for the specific language governing permissions and limitations
19 # For those usages not covered by the Apache License, Version 2.0 please
20 # contact: helena.mcgough@intel.com or adrian.hoban@intel.com
22 """Carry out alarming requests via Aodh API."""
27 from core
.message_bus
.producer
import KafkaProducer
29 from kafka
import KafkaConsumer
31 from plugins
.OpenStack
.common
import Common
32 from plugins
.OpenStack
.response
import OpenStack_Response
34 __author__
= "Helena McGough"
37 "average_memory_usage_above_threshold": "average_memory_utilization",
38 "disk_read_ops": "disk_read_ops",
39 "disk_write_ops": "disk_write_ops",
40 "disk_read_bytes": "disk_read_bytes",
41 "disk_write_bytes": "disk_write_bytes",
42 "net_packets_dropped": "packets_dropped",
43 "packets_in_above_threshold": "packets_received",
44 "packets_out_above_threshold": "packets_sent",
45 "cpu_utilization_above_threshold": "cpu_utilization"}
51 "critical": "critical",
52 "indeterminate": "critical"}
62 class Alarming(object):
63 """Carries out alarming requests and responses via Aodh API."""
66 """Create the OpenStack alarming instance."""
67 self
._common
= Common()
69 # TODO(mcgoughh): Remove hardcoded kafkaconsumer
70 # Initialize a generic consumer object to consume message from the SO
71 server
= {'server': 'localhost:9092', 'topic': 'alarm_request'}
72 self
._consumer
= KafkaConsumer(server
['topic'],
74 bootstrap_servers
=server
['server'])
76 # Use the Response class to generate valid json response messages
77 self
._response
= OpenStack_Response()
79 # Initializer a producer to send responses back to SO
80 self
._producer
= KafkaProducer("alarm_response")
83 """Consume info from the message bus to manage alarms."""
84 # Check the alarming functionlity that needs to be performed
85 for message
in self
._consumer
:
87 values
= json
.loads(message
.value
)
88 vim_type
= values
['vim_type'].lower()
90 if vim_type
== "openstack":
91 log
.info("Alarm action required: %s" % (message
.topic
))
93 # Generate and auth_token and endpoint for request
94 auth_token
, endpoint
= self
.authenticate()
96 if message
.key
== "create_alarm_request":
97 # Configure/Update an alarm
98 alarm_details
= values
['alarm_create_request']
100 alarm_id
, alarm_status
= self
.configure_alarm(
101 endpoint
, auth_token
, alarm_details
)
103 # Generate a valid response message, send via producer
105 resp_message
= self
._response
.generate_response(
106 'create_alarm_response', status
=alarm_status
,
108 cor_id
=alarm_details
['correlation_id'])
109 self
._producer
.create_alarm_response(
110 'create_alarm_resonse', resp_message
,
112 except Exception as exc
:
113 log
.warn("Response creation failed: %s", exc
)
115 elif message
.key
== "list_alarm_request":
116 # Check for a specifed: alarm_name, resource_uuid, severity
117 # and generate the appropriate list
118 list_details
= values
['alarm_list_request']
120 alarm_list
= self
.list_alarms(
121 endpoint
, auth_token
, list_details
)
124 # Generate and send a list response back
125 resp_message
= self
._response
.generate_response(
126 'list_alarm_response', alarm_list
=alarm_list
,
127 cor_id
=list_details
['correlation_id'])
128 self
._producer
.list_alarm_response(
129 'list_alarm_response', resp_message
,
131 except Exception as exc
:
132 log
.warn("Failed to send a valid response back.")
134 elif message
.key
== "delete_alarm_request":
135 request_details
= values
['alarm_delete_request']
136 alarm_id
= request_details
['alarm_uuid']
138 resp_status
= self
.delete_alarm(
139 endpoint
, auth_token
, alarm_id
)
141 # Generate and send a response message
143 resp_message
= self
._response
.generate_response(
144 'delete_alarm_response', alarm_id
=alarm_id
,
146 cor_id
=request_details
['correlation_id'])
147 self
._producer
.delete_alarm_response(
148 'delete_alarm_response', resp_message
,
150 except Exception as exc
:
151 log
.warn("Failed to create delete reponse:%s", exc
)
153 elif message
.key
== "acknowledge_alarm":
154 # Acknowledge that an alarm has been dealt with by the SO
155 alarm_id
= values
['ack_details']['alarm_uuid']
157 response
= self
.update_alarm_state(
158 endpoint
, auth_token
, alarm_id
)
160 # Log if an alarm was reset
162 log
.info("Acknowledged the alarm and cleared it.")
164 log
.warn("Failed to acknowledge/clear the alarm.")
166 elif message
.key
== "update_alarm_request":
167 # Update alarm configurations
168 alarm_details
= values
['alarm_update_request']
170 alarm_id
, status
= self
.update_alarm(
171 endpoint
, auth_token
, alarm_details
)
173 # Generate a response for an update request
175 resp_message
= self
._response
.generate_response(
176 'update_alarm_response', alarm_id
=alarm_id
,
177 cor_id
=alarm_details
['correlation_id'],
179 self
._producer
.update_alarm_response(
180 'update_alarm_response', resp_message
,
182 except Exception as exc
:
183 log
.warn("Failed to send an update response:%s", exc
)
186 log
.debug("Unknown key, no action will be performed")
188 log
.info("Message topic not relevant to this plugin: %s",
193 def configure_alarm(self
, endpoint
, auth_token
, values
):
194 """Create requested alarm in Aodh."""
195 url
= "{}/v2/alarms/".format(endpoint
)
197 # Check if the desired alarm is supported
198 alarm_name
= values
['alarm_name'].lower()
199 metric_name
= values
['metric_name'].lower()
200 resource_id
= values
['resource_uuid']
202 if alarm_name
not in ALARM_NAMES
.keys():
203 log
.warn("This alarm is not supported, by a valid metric.")
205 if ALARM_NAMES
[alarm_name
] != metric_name
:
206 log
.warn("This is not the correct metric for this alarm.")
209 # Check for the required metric
210 metric_id
= self
.check_for_metric(auth_token
, metric_name
, resource_id
)
213 if metric_id
is not None:
214 # Create the alarm if metric is available
215 payload
= self
.check_payload(values
, metric_name
, resource_id
,
217 new_alarm
= self
._common
._perform
_request
(
218 url
, auth_token
, req_type
="post", payload
=payload
)
219 return json
.loads(new_alarm
.text
)['alarm_id'], True
221 log
.warn("The required Gnocchi metric does not exist.")
224 except Exception as exc
:
225 log
.warn("Failed to create the alarm: %s", exc
)
228 def delete_alarm(self
, endpoint
, auth_token
, alarm_id
):
229 """Delete alarm function."""
230 url
= "{}/v2/alarms/%s".format(endpoint
) % (alarm_id
)
233 result
= self
._common
._perform
_request
(
234 url
, auth_token
, req_type
="delete")
235 if str(result
.status_code
) == "404":
236 # If status code is 404 alarm did not exist
241 except Exception as exc
:
242 log
.warn("Failed to delete alarm: %s because %s.", alarm_id
, exc
)
245 def list_alarms(self
, endpoint
, auth_token
, list_details
):
246 """Generate the requested list of alarms."""
247 url
= "{}/v2/alarms/".format(endpoint
)
248 a_list
, name_list
, sev_list
, res_list
= [], [], [], []
250 # TODO(mcgoughh): for now resource_id is a mandatory field
251 resource
= list_details
['resource_uuid']
253 # Checking what fields are specified for a list request
255 name
= list_details
['alarm_name'].lower()
256 if name
not in ALARM_NAMES
.keys():
257 log
.warn("This alarm is not supported, won't be used!")
259 except KeyError as exc
:
260 log
.info("Alarm name isn't specified.")
264 severity
= list_details
['severity'].lower()
265 sev
= SEVERITIES
[severity
]
266 except KeyError as exc
:
267 log
.info("Severity is unspecified/incorrectly configured")
270 # Perform the request to get the desired list
272 result
= self
._common
._perform
_request
(
273 url
, auth_token
, req_type
="get")
275 if result
is not None:
276 # Get list based on resource id
277 for alarm
in json
.loads(result
.text
):
278 rule
= alarm
['gnocchi_resources_threshold_rule']
279 if resource
== rule
['resource_id']:
280 res_list
.append(str(alarm
))
282 log
.info("No alarms for this resource")
285 # Generate specified listed if requested
286 if name
is not None and sev
is not None:
287 log
.info("Return a list of %s alarms with %s severity.",
289 for alarm
in json
.loads(result
.text
):
290 if name
== alarm
['name']:
291 name_list
.append(str(alarm
))
292 for alarm
in json
.loads(result
.text
):
293 if sev
== alarm
['severity']:
294 sev_list
.append(str(alarm
))
295 name_sev_list
= list(set(name_list
).intersection(sev_list
))
296 a_list
= list(set(name_sev_list
).intersection(res_list
))
297 elif name
is not None:
298 log
.info("Returning a %s list of alarms.", name
)
299 for alarm
in json
.loads(result
.text
):
300 if name
== alarm
['name']:
301 name_list
.append(str(alarm
))
302 a_list
= list(set(name_list
).intersection(res_list
))
303 elif sev
is not None:
304 log
.info("Returning %s severity alarm list.", sev
)
305 for alarm
in json
.loads(result
.text
):
306 if sev
== alarm
['severity']:
307 sev_list
.append(str(alarm
))
308 a_list
= list(set(sev_list
).intersection(res_list
))
310 log
.info("Returning an entire list of alarms.")
313 log
.info("There are no alarms!")
315 except Exception as exc
:
316 log
.info("Failed to generate required list: %s", exc
)
321 def update_alarm_state(self
, endpoint
, auth_token
, alarm_id
):
322 """Set the state of an alarm to ok when ack message is received."""
323 url
= "{}/v2/alarms/%s/state".format(endpoint
) % alarm_id
324 payload
= json
.dumps("ok")
327 self
._common
._perform
_request
(
328 url
, auth_token
, req_type
="put", payload
=payload
)
330 except Exception as exc
:
331 log
.warn("Unable to update alarm state: %s", exc
)
334 def update_alarm(self
, endpoint
, auth_token
, values
):
335 """Get alarm name for an alarm configuration update."""
336 # Get already existing alarm details
337 url
= "{}/v2/alarms/%s".format(endpoint
) % values
['alarm_uuid']
339 # Gets current configurations about the alarm
341 result
= self
._common
._perform
_request
(
342 url
, auth_token
, req_type
="get")
343 alarm_name
= json
.loads(result
.text
)['name']
344 rule
= json
.loads(result
.text
)['gnocchi_resources_threshold_rule']
345 alarm_state
= json
.loads(result
.text
)['state']
346 resource_id
= rule
['resource_id']
347 metric_name
= rule
['metric']
348 except Exception as exc
:
349 log
.warn("Failed to retreive existing alarm info: %s.\
350 Can only update OSM alarms.", exc
)
353 # Generates and check payload configuration for alarm update
354 payload
= self
.check_payload(values
, metric_name
, resource_id
,
355 alarm_name
, alarm_state
=alarm_state
)
357 # Updates the alarm configurations with the valid payload
358 if payload
is not None:
360 update_alarm
= self
._common
._perform
_request
(
361 url
, auth_token
, req_type
="put", payload
=payload
)
363 return json
.loads(update_alarm
.text
)['alarm_id'], True
364 except Exception as exc
:
365 log
.warn("Alarm update could not be performed: %s", exc
)
369 def check_payload(self
, values
, metric_name
, resource_id
,
370 alarm_name
, alarm_state
=None):
371 """Check that the payload is configuration for update/create alarm."""
373 # Check state and severity
374 severity
= values
['severity'].lower()
375 if severity
== "indeterminate":
376 alarm_state
= "insufficient data"
377 if alarm_state
is None:
380 statistic
= values
['statistic'].lower()
381 # Try to configure the payload for the update/create request
382 # Can only update: threshold, operation, statistic and
383 # the severity of the alarm
384 rule
= {'threshold': values
['threshold_value'],
385 'comparison_operator': values
['operation'].lower(),
386 'metric': metric_name
,
387 'resource_id': resource_id
,
388 'resource_type': 'generic',
389 'aggregation_method': STATISTICS
[statistic
]}
390 payload
= json
.dumps({'state': alarm_state
,
392 'severity': SEVERITIES
[severity
],
393 'type': 'gnocchi_resources_threshold',
394 'gnocchi_resources_threshold_rule': rule
, })
396 except KeyError as exc
:
397 log
.warn("Alarm is not configured correctly: %s", exc
)
400 def authenticate(self
):
401 """Generate an authentication token and endpoint for alarm request."""
403 # Check for a tenant_id
404 auth_token
= self
._common
._authenticate
()
405 endpoint
= self
._common
.get_endpoint("alarming")
406 return auth_token
, endpoint
407 except Exception as exc
:
408 log
.warn("Authentication to Keystone failed:%s", exc
)
411 def get_alarm_state(self
, endpoint
, auth_token
, alarm_id
):
412 """Get the state of the alarm."""
413 url
= "{}/v2/alarms/%s/state".format(endpoint
) % alarm_id
416 alarm_state
= self
._common
._perform
_request
(
417 url
, auth_token
, req_type
="get")
418 return json
.loads(alarm_state
.text
)
419 except Exception as exc
:
420 log
.warn("Failed to get the state of the alarm:%s", exc
)
423 def check_for_metric(self
, auth_token
, m_name
, r_id
):
424 """Check for the alarm metric."""
426 endpoint
= self
._common
.get_endpoint("metric")
428 url
= "{}/v1/metric/".format(endpoint
)
429 metric_list
= self
._common
._perform
_request
(
430 url
, auth_token
, req_type
="get")
432 for metric
in json
.loads(metric_list
.text
):
433 name
= metric
['name']
434 resource
= metric
['resource_id']
435 if (name
== m_name
and resource
== r_id
):
436 metric_id
= metric
['id']
437 log
.info("The required metric exists, an alarm will be created.")
439 except Exception as exc
:
440 log
.info("Desired Gnocchi metric not found:%s", exc
)