Show More
@@ -1,662 +1,672 b'' | |||
|
1 | 1 | # -*- coding: utf-8 -*- |
|
2 | 2 | |
|
3 | 3 | # Copyright 2010 - 2017 RhodeCode GmbH and the AppEnlight project authors |
|
4 | 4 | # |
|
5 | 5 | # Licensed under the Apache License, Version 2.0 (the "License"); |
|
6 | 6 | # you may not use this file except in compliance with the License. |
|
7 | 7 | # You may obtain a copy of the License at |
|
8 | 8 | # |
|
9 | 9 | # http://www.apache.org/licenses/LICENSE-2.0 |
|
10 | 10 | # |
|
11 | 11 | # Unless required by applicable law or agreed to in writing, software |
|
12 | 12 | # distributed under the License is distributed on an "AS IS" BASIS, |
|
13 | 13 | # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
|
14 | 14 | # See the License for the specific language governing permissions and |
|
15 | 15 | # limitations under the License. |
|
16 | 16 | |
|
17 | 17 | import bisect |
|
18 | 18 | import collections |
|
19 | 19 | import math |
|
20 | 20 | from datetime import datetime, timedelta |
|
21 | 21 | |
|
22 | 22 | import sqlalchemy as sa |
|
23 | 23 | import pyelasticsearch |
|
24 | 24 | |
|
25 | 25 | from celery.utils.log import get_task_logger |
|
26 | 26 | from zope.sqlalchemy import mark_changed |
|
27 | 27 | from pyramid.threadlocal import get_current_request, get_current_registry |
|
28 | 28 | from ziggurat_foundations.models.services.resource import ResourceService |
|
29 | 29 | |
|
30 | 30 | from appenlight.celery import celery |
|
31 | 31 | from appenlight.models.report_group import ReportGroup |
|
32 | 32 | from appenlight.models import DBSession, Datastores |
|
33 | 33 | from appenlight.models.report import Report |
|
34 | 34 | from appenlight.models.log import Log |
|
35 | 35 | from appenlight.models.metric import Metric |
|
36 | 36 | from appenlight.models.event import Event |
|
37 | 37 | |
|
38 | 38 | from appenlight.models.services.application import ApplicationService |
|
39 | 39 | from appenlight.models.services.event import EventService |
|
40 | 40 | from appenlight.models.services.log import LogService |
|
41 | 41 | from appenlight.models.services.report import ReportService |
|
42 | 42 | from appenlight.models.services.report_group import ReportGroupService |
|
43 | 43 | from appenlight.models.services.user import UserService |
|
44 | 44 | from appenlight.models.tag import Tag |
|
45 | 45 | from appenlight.lib import print_traceback |
|
46 | 46 | from appenlight.lib.utils import parse_proto, in_batches |
|
47 | 47 | from appenlight.lib.ext_json import json |
|
48 | 48 | from appenlight.lib.redis_keys import REDIS_KEYS |
|
49 | 49 | from appenlight.lib.enums import ReportType |
|
50 | 50 | |
|
51 | 51 | log = get_task_logger(__name__) |
|
52 | 52 | |
|
53 | 53 | sample_boundries = list(range(100, 1000, 100)) + \ |
|
54 | 54 | list(range(1000, 10000, 1000)) + \ |
|
55 | 55 | list(range(10000, 100000, 5000)) |
|
56 | 56 | |
|
57 | 57 | |
|
58 | 58 | def pick_sample(total_occurences, report_type=None): |
|
59 | 59 | every = 1.0 |
|
60 | 60 | position = bisect.bisect_left(sample_boundries, total_occurences) |
|
61 | 61 | if position > 0: |
|
62 | 62 | if report_type == ReportType.not_found: |
|
63 | 63 | divide = 10.0 |
|
64 | 64 | else: |
|
65 | 65 | divide = 100.0 |
|
66 | 66 | every = sample_boundries[position - 1] / divide |
|
67 | 67 | return total_occurences % every == 0 |
|
68 | 68 | |
|
69 | 69 | |
|
70 | 70 | @celery.task(queue="default", default_retry_delay=1, max_retries=2) |
|
71 | 71 | def test_exception_task(): |
|
72 | 72 | log.error('test celery log', extra={'location': 'celery'}) |
|
73 | 73 | log.warning('test celery log', extra={'location': 'celery'}) |
|
74 | 74 | raise Exception('Celery exception test') |
|
75 | 75 | |
|
76 | 76 | |
|
77 | 77 | @celery.task(queue="default", default_retry_delay=1, max_retries=2) |
|
78 | 78 | def test_retry_exception_task(): |
|
79 | 79 | try: |
|
80 | 80 | import time |
|
81 | 81 | |
|
82 | 82 | time.sleep(1.3) |
|
83 | 83 | log.error('test retry celery log', extra={'location': 'celery'}) |
|
84 | 84 | log.warning('test retry celery log', extra={'location': 'celery'}) |
|
85 | 85 | raise Exception('Celery exception test') |
|
86 | 86 | except Exception as exc: |
|
87 | if celery.conf["CELERY_EAGER_PROPAGATES_EXCEPTIONS"]: | |
|
88 | raise | |
|
87 | 89 | test_retry_exception_task.retry(exc=exc) |
|
88 | 90 | |
|
89 | 91 | |
|
90 | 92 | @celery.task(queue="reports", default_retry_delay=600, max_retries=144) |
|
91 | 93 | def add_reports(resource_id, request_params, dataset, **kwargs): |
|
92 | 94 | proto_version = parse_proto(request_params.get('protocol_version', '')) |
|
93 | 95 | current_time = datetime.utcnow().replace(second=0, microsecond=0) |
|
94 | 96 | try: |
|
95 | 97 | # we will store solr docs here for single insert |
|
96 | 98 | es_report_docs = {} |
|
97 | 99 | es_report_group_docs = {} |
|
98 | 100 | resource = ApplicationService.by_id(resource_id) |
|
99 | 101 | |
|
100 | 102 | tags = [] |
|
101 | 103 | es_slow_calls_docs = {} |
|
102 | 104 | es_reports_stats_rows = {} |
|
103 | 105 | for report_data in dataset: |
|
104 | 106 | # build report details for later |
|
105 | 107 | added_details = 0 |
|
106 | 108 | report = Report() |
|
107 | 109 | report.set_data(report_data, resource, proto_version) |
|
108 | 110 | report._skip_ft_index = True |
|
109 | 111 | |
|
110 | 112 | # find latest group in this months partition |
|
111 | 113 | report_group = ReportGroupService.by_hash_and_resource( |
|
112 | 114 | report.resource_id, |
|
113 | 115 | report.grouping_hash, |
|
114 | 116 | since_when=datetime.utcnow().date().replace(day=1) |
|
115 | 117 | ) |
|
116 | 118 | occurences = report_data.get('occurences', 1) |
|
117 | 119 | if not report_group: |
|
118 | 120 | # total reports will be +1 moment later |
|
119 | 121 | report_group = ReportGroup(grouping_hash=report.grouping_hash, |
|
120 | 122 | occurences=0, total_reports=0, |
|
121 | 123 | last_report=0, |
|
122 | 124 | priority=report.priority, |
|
123 | 125 | error=report.error, |
|
124 | 126 | first_timestamp=report.start_time) |
|
125 | 127 | report_group._skip_ft_index = True |
|
126 | 128 | report_group.report_type = report.report_type |
|
127 | 129 | report.report_group_time = report_group.first_timestamp |
|
128 | 130 | add_sample = pick_sample(report_group.occurences, |
|
129 | 131 | report_type=report_group.report_type) |
|
130 | 132 | if add_sample: |
|
131 | 133 | resource.report_groups.append(report_group) |
|
132 | 134 | report_group.reports.append(report) |
|
133 | 135 | added_details += 1 |
|
134 | 136 | DBSession.flush() |
|
135 | 137 | if report.partition_id not in es_report_docs: |
|
136 | 138 | es_report_docs[report.partition_id] = [] |
|
137 | 139 | es_report_docs[report.partition_id].append(report.es_doc()) |
|
138 | 140 | tags.extend(list(report.tags.items())) |
|
139 | 141 | slow_calls = report.add_slow_calls(report_data, report_group) |
|
140 | 142 | DBSession.flush() |
|
141 | 143 | for s_call in slow_calls: |
|
142 | 144 | if s_call.partition_id not in es_slow_calls_docs: |
|
143 | 145 | es_slow_calls_docs[s_call.partition_id] = [] |
|
144 | 146 | es_slow_calls_docs[s_call.partition_id].append( |
|
145 | 147 | s_call.es_doc()) |
|
146 | 148 | # try generating new stat rows if needed |
|
147 | 149 | else: |
|
148 | 150 | # required for postprocessing to not fail later |
|
149 | 151 | report.report_group = report_group |
|
150 | 152 | |
|
151 | 153 | stat_row = ReportService.generate_stat_rows( |
|
152 | 154 | report, resource, report_group) |
|
153 | 155 | if stat_row.partition_id not in es_reports_stats_rows: |
|
154 | 156 | es_reports_stats_rows[stat_row.partition_id] = [] |
|
155 | 157 | es_reports_stats_rows[stat_row.partition_id].append( |
|
156 | 158 | stat_row.es_doc()) |
|
157 | 159 | |
|
158 | 160 | # see if we should mark 10th occurence of report |
|
159 | 161 | last_occurences_10 = int(math.floor(report_group.occurences / 10)) |
|
160 | 162 | curr_occurences_10 = int(math.floor( |
|
161 | 163 | (report_group.occurences + report.occurences) / 10)) |
|
162 | 164 | last_occurences_100 = int( |
|
163 | 165 | math.floor(report_group.occurences / 100)) |
|
164 | 166 | curr_occurences_100 = int(math.floor( |
|
165 | 167 | (report_group.occurences + report.occurences) / 100)) |
|
166 | 168 | notify_occurences_10 = last_occurences_10 != curr_occurences_10 |
|
167 | 169 | notify_occurences_100 = last_occurences_100 != curr_occurences_100 |
|
168 | 170 | report_group.occurences = ReportGroup.occurences + occurences |
|
169 | 171 | report_group.last_timestamp = report.start_time |
|
170 | 172 | report_group.summed_duration = ReportGroup.summed_duration + report.duration |
|
171 | 173 | summed_duration = ReportGroup.summed_duration + report.duration |
|
172 | 174 | summed_occurences = ReportGroup.occurences + occurences |
|
173 | 175 | report_group.average_duration = summed_duration / summed_occurences |
|
174 | 176 | report_group.run_postprocessing(report) |
|
175 | 177 | if added_details: |
|
176 | 178 | report_group.total_reports = ReportGroup.total_reports + 1 |
|
177 | 179 | report_group.last_report = report.id |
|
178 | 180 | report_group.set_notification_info(notify_10=notify_occurences_10, |
|
179 | 181 | notify_100=notify_occurences_100) |
|
180 | 182 | DBSession.flush() |
|
181 | 183 | report_group.get_report().notify_channel(report_group) |
|
182 | 184 | if report_group.partition_id not in es_report_group_docs: |
|
183 | 185 | es_report_group_docs[report_group.partition_id] = [] |
|
184 | 186 | es_report_group_docs[report_group.partition_id].append( |
|
185 | 187 | report_group.es_doc()) |
|
186 | 188 | |
|
187 | 189 | action = 'REPORT' |
|
188 | 190 | log_msg = '%s: %s %s, client: %s, proto: %s' % ( |
|
189 | 191 | action, |
|
190 | 192 | report_data.get('http_status', 'unknown'), |
|
191 | 193 | str(resource), |
|
192 | 194 | report_data.get('client'), |
|
193 | 195 | proto_version) |
|
194 | 196 | log.info(log_msg) |
|
195 | 197 | total_reports = len(dataset) |
|
196 | 198 | redis_pipeline = Datastores.redis.pipeline(transaction=False) |
|
197 | 199 | key = REDIS_KEYS['counters']['reports_per_minute'].format(current_time) |
|
198 | 200 | redis_pipeline.incr(key, total_reports) |
|
199 | 201 | redis_pipeline.expire(key, 3600 * 24) |
|
200 | 202 | key = REDIS_KEYS['counters']['events_per_minute_per_user'].format( |
|
201 | 203 | resource.owner_user_id, current_time) |
|
202 | 204 | redis_pipeline.incr(key, total_reports) |
|
203 | 205 | redis_pipeline.expire(key, 3600) |
|
204 | 206 | key = REDIS_KEYS['counters']['reports_per_hour_per_app'].format( |
|
205 | 207 | resource_id, current_time.replace(minute=0)) |
|
206 | 208 | redis_pipeline.incr(key, total_reports) |
|
207 | 209 | redis_pipeline.expire(key, 3600 * 24 * 7) |
|
208 | 210 | redis_pipeline.sadd( |
|
209 | 211 | REDIS_KEYS['apps_that_got_new_data_per_hour'].format( |
|
210 | 212 | current_time.replace(minute=0)), resource_id) |
|
211 | 213 | redis_pipeline.execute() |
|
212 | 214 | |
|
213 | 215 | add_reports_es(es_report_group_docs, es_report_docs) |
|
214 | 216 | add_reports_slow_calls_es(es_slow_calls_docs) |
|
215 | 217 | add_reports_stats_rows_es(es_reports_stats_rows) |
|
216 | 218 | return True |
|
217 | 219 | except Exception as exc: |
|
218 | 220 | print_traceback(log) |
|
221 | if celery.conf["CELERY_EAGER_PROPAGATES_EXCEPTIONS"]: | |
|
222 | raise | |
|
219 | 223 | add_reports.retry(exc=exc) |
|
220 | 224 | |
|
221 | 225 | |
|
222 | 226 | @celery.task(queue="es", default_retry_delay=600, max_retries=144) |
|
223 | 227 | def add_reports_es(report_group_docs, report_docs): |
|
224 | 228 | for k, v in report_group_docs.items(): |
|
225 | 229 | Datastores.es.bulk_index(k, 'report_group', v, id_field="_id") |
|
226 | 230 | for k, v in report_docs.items(): |
|
227 | 231 | Datastores.es.bulk_index(k, 'report', v, id_field="_id", |
|
228 | 232 | parent_field='_parent') |
|
229 | 233 | |
|
230 | 234 | |
|
231 | 235 | @celery.task(queue="es", default_retry_delay=600, max_retries=144) |
|
232 | 236 | def add_reports_slow_calls_es(es_docs): |
|
233 | 237 | for k, v in es_docs.items(): |
|
234 | 238 | Datastores.es.bulk_index(k, 'log', v) |
|
235 | 239 | |
|
236 | 240 | |
|
237 | 241 | @celery.task(queue="es", default_retry_delay=600, max_retries=144) |
|
238 | 242 | def add_reports_stats_rows_es(es_docs): |
|
239 | 243 | for k, v in es_docs.items(): |
|
240 | 244 | Datastores.es.bulk_index(k, 'log', v) |
|
241 | 245 | |
|
242 | 246 | |
|
243 | 247 | @celery.task(queue="logs", default_retry_delay=600, max_retries=144) |
|
244 | 248 | def add_logs(resource_id, request_params, dataset, **kwargs): |
|
245 | 249 | proto_version = request_params.get('protocol_version') |
|
246 | 250 | current_time = datetime.utcnow().replace(second=0, microsecond=0) |
|
247 | 251 | |
|
248 | 252 | try: |
|
249 | 253 | es_docs = collections.defaultdict(list) |
|
250 | 254 | resource = ApplicationService.by_id_cached()(resource_id) |
|
251 | 255 | resource = DBSession.merge(resource, load=False) |
|
252 | 256 | ns_pairs = [] |
|
253 | 257 | for entry in dataset: |
|
254 | 258 | # gather pk and ns so we can remove older versions of row later |
|
255 | 259 | if entry['primary_key'] is not None: |
|
256 | 260 | ns_pairs.append({"pk": entry['primary_key'], |
|
257 | 261 | "ns": entry['namespace']}) |
|
258 | 262 | log_entry = Log() |
|
259 | 263 | log_entry.set_data(entry, resource=resource) |
|
260 | 264 | log_entry._skip_ft_index = True |
|
261 | 265 | resource.logs.append(log_entry) |
|
262 | 266 | DBSession.flush() |
|
263 | 267 | # insert non pk rows first |
|
264 | 268 | if entry['primary_key'] is None: |
|
265 | 269 | es_docs[log_entry.partition_id].append(log_entry.es_doc()) |
|
266 | 270 | |
|
267 | 271 | # 2nd pass to delete all log entries from db foe same pk/ns pair |
|
268 | 272 | if ns_pairs: |
|
269 | 273 | ids_to_delete = [] |
|
270 | 274 | es_docs = collections.defaultdict(list) |
|
271 | 275 | es_docs_to_delete = collections.defaultdict(list) |
|
272 | 276 | found_pkey_logs = LogService.query_by_primary_key_and_namespace( |
|
273 | 277 | list_of_pairs=ns_pairs) |
|
274 | 278 | log_dict = {} |
|
275 | 279 | for log_entry in found_pkey_logs: |
|
276 | 280 | log_key = (log_entry.primary_key, log_entry.namespace) |
|
277 | 281 | if log_key not in log_dict: |
|
278 | 282 | log_dict[log_key] = [] |
|
279 | 283 | log_dict[log_key].append(log_entry) |
|
280 | 284 | |
|
281 | 285 | for ns, entry_list in log_dict.items(): |
|
282 | 286 | entry_list = sorted(entry_list, key=lambda x: x.timestamp) |
|
283 | 287 | # newest row needs to be indexed in es |
|
284 | 288 | log_entry = entry_list[-1] |
|
285 | 289 | # delete everything from pg and ES, leave the last row in pg |
|
286 | 290 | for e in entry_list[:-1]: |
|
287 | 291 | ids_to_delete.append(e.log_id) |
|
288 | 292 | es_docs_to_delete[e.partition_id].append(e.delete_hash) |
|
289 | 293 | |
|
290 | 294 | es_docs_to_delete[log_entry.partition_id].append( |
|
291 | 295 | log_entry.delete_hash) |
|
292 | 296 | |
|
293 | 297 | es_docs[log_entry.partition_id].append(log_entry.es_doc()) |
|
294 | 298 | |
|
295 | 299 | if ids_to_delete: |
|
296 | 300 | query = DBSession.query(Log).filter( |
|
297 | 301 | Log.log_id.in_(ids_to_delete)) |
|
298 | 302 | query.delete(synchronize_session=False) |
|
299 | 303 | if es_docs_to_delete: |
|
300 | 304 | # batch this to avoid problems with default ES bulk limits |
|
301 | 305 | for es_index in es_docs_to_delete.keys(): |
|
302 | 306 | for batch in in_batches(es_docs_to_delete[es_index], 20): |
|
303 | 307 | query = {'terms': {'delete_hash': batch}} |
|
304 | 308 | |
|
305 | 309 | try: |
|
306 | 310 | Datastores.es.delete_by_query( |
|
307 | 311 | es_index, 'log', query) |
|
308 | 312 | except pyelasticsearch.ElasticHttpNotFoundError as exc: |
|
309 | 313 | msg = 'skipping index {}'.format(es_index) |
|
310 | 314 | log.info(msg) |
|
311 | 315 | |
|
312 | 316 | total_logs = len(dataset) |
|
313 | 317 | |
|
314 | 318 | log_msg = 'LOG_NEW: %s, entries: %s, proto:%s' % ( |
|
315 | 319 | str(resource), |
|
316 | 320 | total_logs, |
|
317 | 321 | proto_version) |
|
318 | 322 | log.info(log_msg) |
|
319 | 323 | # mark_changed(session) |
|
320 | 324 | redis_pipeline = Datastores.redis.pipeline(transaction=False) |
|
321 | 325 | key = REDIS_KEYS['counters']['logs_per_minute'].format(current_time) |
|
322 | 326 | redis_pipeline.incr(key, total_logs) |
|
323 | 327 | redis_pipeline.expire(key, 3600 * 24) |
|
324 | 328 | key = REDIS_KEYS['counters']['events_per_minute_per_user'].format( |
|
325 | 329 | resource.owner_user_id, current_time) |
|
326 | 330 | redis_pipeline.incr(key, total_logs) |
|
327 | 331 | redis_pipeline.expire(key, 3600) |
|
328 | 332 | key = REDIS_KEYS['counters']['logs_per_hour_per_app'].format( |
|
329 | 333 | resource_id, current_time.replace(minute=0)) |
|
330 | 334 | redis_pipeline.incr(key, total_logs) |
|
331 | 335 | redis_pipeline.expire(key, 3600 * 24 * 7) |
|
332 | 336 | redis_pipeline.sadd( |
|
333 | 337 | REDIS_KEYS['apps_that_got_new_data_per_hour'].format( |
|
334 | 338 | current_time.replace(minute=0)), resource_id) |
|
335 | 339 | redis_pipeline.execute() |
|
336 | 340 | add_logs_es(es_docs) |
|
337 | 341 | return True |
|
338 | 342 | except Exception as exc: |
|
339 | 343 | print_traceback(log) |
|
344 | if celery.conf["CELERY_EAGER_PROPAGATES_EXCEPTIONS"]: | |
|
345 | raise | |
|
340 | 346 | add_logs.retry(exc=exc) |
|
341 | 347 | |
|
342 | 348 | |
|
343 | 349 | @celery.task(queue="es", default_retry_delay=600, max_retries=144) |
|
344 | 350 | def add_logs_es(es_docs): |
|
345 | 351 | for k, v in es_docs.items(): |
|
346 | 352 | Datastores.es.bulk_index(k, 'log', v) |
|
347 | 353 | |
|
348 | 354 | |
|
349 | 355 | @celery.task(queue="metrics", default_retry_delay=600, max_retries=144) |
|
350 | 356 | def add_metrics(resource_id, request_params, dataset, proto_version): |
|
351 | 357 | current_time = datetime.utcnow().replace(second=0, microsecond=0) |
|
352 | 358 | try: |
|
353 | 359 | resource = ApplicationService.by_id_cached()(resource_id) |
|
354 | 360 | resource = DBSession.merge(resource, load=False) |
|
355 | 361 | es_docs = [] |
|
356 | 362 | rows = [] |
|
357 | 363 | for metric in dataset: |
|
358 | 364 | tags = dict(metric['tags']) |
|
359 | 365 | server_n = tags.get('server_name', metric['server_name']).lower() |
|
360 | 366 | tags['server_name'] = server_n or 'unknown' |
|
361 | 367 | new_metric = Metric( |
|
362 | 368 | timestamp=metric['timestamp'], |
|
363 | 369 | resource_id=resource.resource_id, |
|
364 | 370 | namespace=metric['namespace'], |
|
365 | 371 | tags=tags) |
|
366 | 372 | rows.append(new_metric) |
|
367 | 373 | es_docs.append(new_metric.es_doc()) |
|
368 | 374 | session = DBSession() |
|
369 | 375 | session.bulk_save_objects(rows) |
|
370 | 376 | session.flush() |
|
371 | 377 | |
|
372 | 378 | action = 'METRICS' |
|
373 | 379 | metrics_msg = '%s: %s, metrics: %s, proto:%s' % ( |
|
374 | 380 | action, |
|
375 | 381 | str(resource), |
|
376 | 382 | len(dataset), |
|
377 | 383 | proto_version |
|
378 | 384 | ) |
|
379 | 385 | log.info(metrics_msg) |
|
380 | 386 | |
|
381 | 387 | mark_changed(session) |
|
382 | 388 | redis_pipeline = Datastores.redis.pipeline(transaction=False) |
|
383 | 389 | key = REDIS_KEYS['counters']['metrics_per_minute'].format(current_time) |
|
384 | 390 | redis_pipeline.incr(key, len(rows)) |
|
385 | 391 | redis_pipeline.expire(key, 3600 * 24) |
|
386 | 392 | key = REDIS_KEYS['counters']['events_per_minute_per_user'].format( |
|
387 | 393 | resource.owner_user_id, current_time) |
|
388 | 394 | redis_pipeline.incr(key, len(rows)) |
|
389 | 395 | redis_pipeline.expire(key, 3600) |
|
390 | 396 | key = REDIS_KEYS['counters']['metrics_per_hour_per_app'].format( |
|
391 | 397 | resource_id, current_time.replace(minute=0)) |
|
392 | 398 | redis_pipeline.incr(key, len(rows)) |
|
393 | 399 | redis_pipeline.expire(key, 3600 * 24 * 7) |
|
394 | 400 | redis_pipeline.sadd( |
|
395 | 401 | REDIS_KEYS['apps_that_got_new_data_per_hour'].format( |
|
396 | 402 | current_time.replace(minute=0)), resource_id) |
|
397 | 403 | redis_pipeline.execute() |
|
398 | 404 | add_metrics_es(es_docs) |
|
399 | 405 | return True |
|
400 | 406 | except Exception as exc: |
|
401 | 407 | print_traceback(log) |
|
408 | if celery.conf["CELERY_EAGER_PROPAGATES_EXCEPTIONS"]: | |
|
409 | raise | |
|
402 | 410 | add_metrics.retry(exc=exc) |
|
403 | 411 | |
|
404 | 412 | |
|
405 | 413 | @celery.task(queue="es", default_retry_delay=600, max_retries=144) |
|
406 | 414 | def add_metrics_es(es_docs): |
|
407 | 415 | for doc in es_docs: |
|
408 | 416 | partition = 'rcae_m_%s' % doc['timestamp'].strftime('%Y_%m_%d') |
|
409 | 417 | Datastores.es.index(partition, 'log', doc) |
|
410 | 418 | |
|
411 | 419 | |
|
412 | 420 | @celery.task(queue="default", default_retry_delay=5, max_retries=2) |
|
413 | 421 | def check_user_report_notifications(resource_id): |
|
414 | 422 | since_when = datetime.utcnow() |
|
415 | 423 | try: |
|
416 | 424 | request = get_current_request() |
|
417 | 425 | application = ApplicationService.by_id(resource_id) |
|
418 | 426 | if not application: |
|
419 | 427 | return |
|
420 | 428 | error_key = REDIS_KEYS['reports_to_notify_per_type_per_app'].format( |
|
421 | 429 | ReportType.error, resource_id) |
|
422 | 430 | slow_key = REDIS_KEYS['reports_to_notify_per_type_per_app'].format( |
|
423 | 431 | ReportType.slow, resource_id) |
|
424 | 432 | error_group_ids = Datastores.redis.smembers(error_key) |
|
425 | 433 | slow_group_ids = Datastores.redis.smembers(slow_key) |
|
426 | 434 | Datastores.redis.delete(error_key) |
|
427 | 435 | Datastores.redis.delete(slow_key) |
|
428 | 436 | err_gids = [int(g_id) for g_id in error_group_ids] |
|
429 | 437 | slow_gids = [int(g_id) for g_id in list(slow_group_ids)] |
|
430 | 438 | group_ids = err_gids + slow_gids |
|
431 | 439 | occurence_dict = {} |
|
432 | 440 | for g_id in group_ids: |
|
433 | 441 | key = REDIS_KEYS['counters']['report_group_occurences'].format( |
|
434 | 442 | g_id) |
|
435 | 443 | val = Datastores.redis.get(key) |
|
436 | 444 | Datastores.redis.delete(key) |
|
437 | 445 | if val: |
|
438 | 446 | occurence_dict[g_id] = int(val) |
|
439 | 447 | else: |
|
440 | 448 | occurence_dict[g_id] = 1 |
|
441 | 449 | report_groups = ReportGroupService.by_ids(group_ids) |
|
442 | 450 | report_groups.options(sa.orm.joinedload(ReportGroup.last_report_ref)) |
|
443 | 451 | |
|
444 | 452 | ApplicationService.check_for_groups_alert( |
|
445 | 453 | application, 'alert', report_groups=report_groups, |
|
446 | 454 | occurence_dict=occurence_dict) |
|
447 | 455 | users = set([p.user for p in ResourceService.users_for_perm(application, 'view')]) |
|
448 | 456 | report_groups = report_groups.all() |
|
449 | 457 | for user in users: |
|
450 | 458 | UserService.report_notify(user, request, application, |
|
451 | 459 | report_groups=report_groups, |
|
452 | 460 | occurence_dict=occurence_dict) |
|
453 | 461 | for group in report_groups: |
|
454 | 462 | # marks report_groups as notified |
|
455 | 463 | if not group.notified: |
|
456 | 464 | group.notified = True |
|
457 | 465 | except Exception as exc: |
|
458 | 466 | print_traceback(log) |
|
459 | 467 | raise |
|
460 | 468 | |
|
461 | 469 | |
|
462 | 470 | @celery.task(queue="default", default_retry_delay=5, max_retries=2) |
|
463 | 471 | def check_alerts(resource_id): |
|
464 | 472 | since_when = datetime.utcnow() |
|
465 | 473 | try: |
|
466 | 474 | request = get_current_request() |
|
467 | 475 | application = ApplicationService.by_id(resource_id) |
|
468 | 476 | if not application: |
|
469 | 477 | return |
|
470 | 478 | error_key = REDIS_KEYS[ |
|
471 | 479 | 'reports_to_notify_per_type_per_app_alerting'].format( |
|
472 | 480 | ReportType.error, resource_id) |
|
473 | 481 | slow_key = REDIS_KEYS[ |
|
474 | 482 | 'reports_to_notify_per_type_per_app_alerting'].format( |
|
475 | 483 | ReportType.slow, resource_id) |
|
476 | 484 | error_group_ids = Datastores.redis.smembers(error_key) |
|
477 | 485 | slow_group_ids = Datastores.redis.smembers(slow_key) |
|
478 | 486 | Datastores.redis.delete(error_key) |
|
479 | 487 | Datastores.redis.delete(slow_key) |
|
480 | 488 | err_gids = [int(g_id) for g_id in error_group_ids] |
|
481 | 489 | slow_gids = [int(g_id) for g_id in list(slow_group_ids)] |
|
482 | 490 | group_ids = err_gids + slow_gids |
|
483 | 491 | occurence_dict = {} |
|
484 | 492 | for g_id in group_ids: |
|
485 | 493 | key = REDIS_KEYS['counters'][ |
|
486 | 494 | 'report_group_occurences_alerting'].format( |
|
487 | 495 | g_id) |
|
488 | 496 | val = Datastores.redis.get(key) |
|
489 | 497 | Datastores.redis.delete(key) |
|
490 | 498 | if val: |
|
491 | 499 | occurence_dict[g_id] = int(val) |
|
492 | 500 | else: |
|
493 | 501 | occurence_dict[g_id] = 1 |
|
494 | 502 | report_groups = ReportGroupService.by_ids(group_ids) |
|
495 | 503 | report_groups.options(sa.orm.joinedload(ReportGroup.last_report_ref)) |
|
496 | 504 | |
|
497 | 505 | ApplicationService.check_for_groups_alert( |
|
498 | 506 | application, 'alert', report_groups=report_groups, |
|
499 | 507 | occurence_dict=occurence_dict, since_when=since_when) |
|
500 | 508 | except Exception as exc: |
|
501 | 509 | print_traceback(log) |
|
502 | 510 | raise |
|
503 | 511 | |
|
504 | 512 | |
|
505 | 513 | @celery.task(queue="default", default_retry_delay=1, max_retries=2) |
|
506 | 514 | def close_alerts(): |
|
507 | 515 | log.warning('Checking alerts') |
|
508 | 516 | since_when = datetime.utcnow() |
|
509 | 517 | try: |
|
510 | 518 | event_types = [Event.types['error_report_alert'], |
|
511 | 519 | Event.types['slow_report_alert'], ] |
|
512 | 520 | statuses = [Event.statuses['active']] |
|
513 | 521 | # get events older than 5 min |
|
514 | 522 | events = EventService.by_type_and_status( |
|
515 | 523 | event_types, |
|
516 | 524 | statuses, |
|
517 | 525 | older_than=(since_when - timedelta(minutes=5))) |
|
518 | 526 | for event in events: |
|
519 | 527 | # see if we can close them |
|
520 | 528 | event.validate_or_close( |
|
521 | 529 | since_when=(since_when - timedelta(minutes=1))) |
|
522 | 530 | except Exception as exc: |
|
523 | 531 | print_traceback(log) |
|
524 | 532 | raise |
|
525 | 533 | |
|
526 | 534 | |
|
527 | 535 | @celery.task(queue="default", default_retry_delay=600, max_retries=144) |
|
528 | 536 | def update_tag_counter(tag_name, tag_value, count): |
|
529 | 537 | try: |
|
530 | 538 | query = DBSession.query(Tag).filter(Tag.name == tag_name).filter( |
|
531 | 539 | sa.cast(Tag.value, sa.types.TEXT) == sa.cast(json.dumps(tag_value), |
|
532 | 540 | sa.types.TEXT)) |
|
533 | 541 | query.update({'times_seen': Tag.times_seen + count, |
|
534 | 542 | 'last_timestamp': datetime.utcnow()}, |
|
535 | 543 | synchronize_session=False) |
|
536 | 544 | session = DBSession() |
|
537 | 545 | mark_changed(session) |
|
538 | 546 | return True |
|
539 | 547 | except Exception as exc: |
|
540 | 548 | print_traceback(log) |
|
549 | if celery.conf["CELERY_EAGER_PROPAGATES_EXCEPTIONS"]: | |
|
550 | raise | |
|
541 | 551 | update_tag_counter.retry(exc=exc) |
|
542 | 552 | |
|
543 | 553 | |
|
544 | 554 | @celery.task(queue="default") |
|
545 | 555 | def update_tag_counters(): |
|
546 | 556 | """ |
|
547 | 557 | Sets task to update counters for application tags |
|
548 | 558 | """ |
|
549 | 559 | tags = Datastores.redis.lrange(REDIS_KEYS['seen_tag_list'], 0, -1) |
|
550 | 560 | Datastores.redis.delete(REDIS_KEYS['seen_tag_list']) |
|
551 | 561 | c = collections.Counter(tags) |
|
552 | 562 | for t_json, count in c.items(): |
|
553 | 563 | tag_info = json.loads(t_json) |
|
554 | 564 | update_tag_counter.delay(tag_info[0], tag_info[1], count) |
|
555 | 565 | |
|
556 | 566 | |
|
557 | 567 | @celery.task(queue="default") |
|
558 | 568 | def daily_digest(): |
|
559 | 569 | """ |
|
560 | 570 | Sends daily digest with top 50 error reports |
|
561 | 571 | """ |
|
562 | 572 | request = get_current_request() |
|
563 | 573 | apps = Datastores.redis.smembers(REDIS_KEYS['apps_that_had_reports']) |
|
564 | 574 | Datastores.redis.delete(REDIS_KEYS['apps_that_had_reports']) |
|
565 | 575 | since_when = datetime.utcnow() - timedelta(hours=8) |
|
566 | 576 | log.warning('Generating daily digests') |
|
567 | 577 | for resource_id in apps: |
|
568 | 578 | resource_id = resource_id.decode('utf8') |
|
569 | 579 | end_date = datetime.utcnow().replace(microsecond=0, second=0) |
|
570 | 580 | filter_settings = {'resource': [resource_id], |
|
571 | 581 | 'tags': [{'name': 'type', |
|
572 | 582 | 'value': ['error'], 'op': None}], |
|
573 | 583 | 'type': 'error', 'start_date': since_when, |
|
574 | 584 | 'end_date': end_date} |
|
575 | 585 | |
|
576 | 586 | reports = ReportGroupService.get_trending( |
|
577 | 587 | request, filter_settings=filter_settings, limit=50) |
|
578 | 588 | |
|
579 | 589 | application = ApplicationService.by_id(resource_id) |
|
580 | 590 | if application: |
|
581 | 591 | users = set([p.user for p in ResourceService.users_for_perm(application, 'view')]) |
|
582 | 592 | for user in users: |
|
583 | 593 | user.send_digest(request, application, reports=reports, |
|
584 | 594 | since_when=since_when) |
|
585 | 595 | |
|
586 | 596 | |
|
587 | 597 | @celery.task(queue="default") |
|
588 | 598 | def notifications_reports(): |
|
589 | 599 | """ |
|
590 | 600 | Loop that checks redis for info and then issues new tasks to celery to |
|
591 | 601 | issue notifications |
|
592 | 602 | """ |
|
593 | 603 | apps = Datastores.redis.smembers(REDIS_KEYS['apps_that_had_reports']) |
|
594 | 604 | Datastores.redis.delete(REDIS_KEYS['apps_that_had_reports']) |
|
595 | 605 | for app in apps: |
|
596 | 606 | log.warning('Notify for app: %s' % app) |
|
597 | 607 | check_user_report_notifications.delay(app.decode('utf8')) |
|
598 | 608 | |
|
599 | 609 | @celery.task(queue="default") |
|
600 | 610 | def alerting_reports(): |
|
601 | 611 | """ |
|
602 | 612 | Loop that checks redis for info and then issues new tasks to celery to |
|
603 | 613 | perform the following: |
|
604 | 614 | - which applications should have new alerts opened |
|
605 | 615 | """ |
|
606 | 616 | |
|
607 | 617 | apps = Datastores.redis.smembers(REDIS_KEYS['apps_that_had_reports_alerting']) |
|
608 | 618 | Datastores.redis.delete(REDIS_KEYS['apps_that_had_reports_alerting']) |
|
609 | 619 | for app in apps: |
|
610 | 620 | log.warning('Notify for app: %s' % app) |
|
611 | 621 | check_alerts.delay(app.decode('utf8')) |
|
612 | 622 | |
|
613 | 623 | |
|
614 | 624 | @celery.task(queue="default", soft_time_limit=3600 * 4, |
|
615 | 625 | hard_time_limit=3600 * 4, max_retries=144) |
|
616 | 626 | def logs_cleanup(resource_id, filter_settings): |
|
617 | 627 | request = get_current_request() |
|
618 | 628 | request.tm.begin() |
|
619 | 629 | es_query = { |
|
620 | 630 | "_source": False, |
|
621 | 631 | "size": 5000, |
|
622 | 632 | "query": { |
|
623 | 633 | "filtered": { |
|
624 | 634 | "filter": { |
|
625 | 635 | "and": [{"term": {"resource_id": resource_id}}] |
|
626 | 636 | } |
|
627 | 637 | } |
|
628 | 638 | } |
|
629 | 639 | } |
|
630 | 640 | |
|
631 | 641 | query = DBSession.query(Log).filter(Log.resource_id == resource_id) |
|
632 | 642 | if filter_settings['namespace']: |
|
633 | 643 | query = query.filter(Log.namespace == filter_settings['namespace'][0]) |
|
634 | 644 | es_query['query']['filtered']['filter']['and'].append( |
|
635 | 645 | {"term": {"namespace": filter_settings['namespace'][0]}} |
|
636 | 646 | ) |
|
637 | 647 | query.delete(synchronize_session=False) |
|
638 | 648 | request.tm.commit() |
|
639 | 649 | result = request.es_conn.search(es_query, index='rcae_l_*', |
|
640 | 650 | doc_type='log', es_scroll='1m', |
|
641 | 651 | es_search_type='scan') |
|
642 | 652 | scroll_id = result['_scroll_id'] |
|
643 | 653 | while True: |
|
644 | 654 | log.warning('log_cleanup, app:{} ns:{} batch'.format( |
|
645 | 655 | resource_id, |
|
646 | 656 | filter_settings['namespace'] |
|
647 | 657 | )) |
|
648 | 658 | es_docs_to_delete = [] |
|
649 | 659 | result = request.es_conn.send_request( |
|
650 | 660 | 'POST', ['_search', 'scroll'], |
|
651 | 661 | body=scroll_id, query_params={"scroll": '1m'}) |
|
652 | 662 | scroll_id = result['_scroll_id'] |
|
653 | 663 | if not result['hits']['hits']: |
|
654 | 664 | break |
|
655 | 665 | for doc in result['hits']['hits']: |
|
656 | 666 | es_docs_to_delete.append({"id": doc['_id'], |
|
657 | 667 | "index": doc['_index']}) |
|
658 | 668 | |
|
659 | 669 | for batch in in_batches(es_docs_to_delete, 10): |
|
660 | 670 | Datastores.es.bulk([Datastores.es.delete_op(doc_type='log', |
|
661 | 671 | **to_del) |
|
662 | 672 | for to_del in batch]) |
General Comments 0
You need to be logged in to leave comments.
Login now