clue-python-sdk-core 0.0.1__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- clue_python_sdk_core/__init__.py +135 -0
- clue_python_sdk_core/adapters.py +2177 -0
- clue_python_sdk_core/bootstrap.py +166 -0
- clue_python_sdk_core/celery.py +734 -0
- clue_python_sdk_core/client.py +335 -0
- clue_python_sdk_core/contracts.py +224 -0
- clue_python_sdk_core/event_size.py +96 -0
- clue_python_sdk_core/http_extraction.py +146 -0
- clue_python_sdk_core/orm.py +534 -0
- clue_python_sdk_core/otel_bridge.py +164 -0
- clue_python_sdk_core/parameter_snapshot.py +152 -0
- clue_python_sdk_core/privacy.py +124 -0
- clue_python_sdk_core/requests_instrumentation.py +341 -0
- clue_python_sdk_core/resources.py +44 -0
- clue_python_sdk_core/runtime.py +894 -0
- clue_python_sdk_core-0.0.1.dist-info/METADATA +11 -0
- clue_python_sdk_core-0.0.1.dist-info/RECORD +19 -0
- clue_python_sdk_core-0.0.1.dist-info/WHEEL +5 -0
- clue_python_sdk_core-0.0.1.dist-info/top_level.txt +1 -0
|
@@ -0,0 +1,335 @@
|
|
|
1
|
+
from __future__ import annotations
|
|
2
|
+
|
|
3
|
+
import json
|
|
4
|
+
import threading
|
|
5
|
+
import uuid
|
|
6
|
+
from collections.abc import Callable, Mapping, Sequence
|
|
7
|
+
from dataclasses import dataclass
|
|
8
|
+
from urllib.request import Request, urlopen
|
|
9
|
+
|
|
10
|
+
from .adapters import JsonValue, build_backend_envelope
|
|
11
|
+
from .contracts import EVENT_CATEGORY_OUTBOUND_REQUEST, EVENT_CATEGORY_REQUEST
|
|
12
|
+
from .event_size import DEFAULT_MAX_EVENT_BYTES, normalize_event_for_size
|
|
13
|
+
|
|
14
|
+
DEFAULT_MAX_PAYLOAD_BYTES = 256 * 1024
|
|
15
|
+
DEFAULT_FLUSH_THRESHOLD_BYTES = 220 * 1024
|
|
16
|
+
DEFAULT_MAX_BUFFER_BYTES = 2 * 1024 * 1024
|
|
17
|
+
DEFAULT_MAX_BUFFER_EVENTS = 10_000
|
|
18
|
+
DEFAULT_FLUSH_INTERVAL_MS = 180_000
|
|
19
|
+
|
|
20
|
+
|
|
21
|
+
def _json_bytes(value: Mapping[str, JsonValue]) -> bytes:
|
|
22
|
+
return json.dumps(value, ensure_ascii=False, separators=(",", ":")).encode("utf-8")
|
|
23
|
+
|
|
24
|
+
|
|
25
|
+
@dataclass(frozen=True)
|
|
26
|
+
class _BatchMetadata:
|
|
27
|
+
batch_id: str
|
|
28
|
+
idempotency_key: str
|
|
29
|
+
sent_at: str
|
|
30
|
+
|
|
31
|
+
|
|
32
|
+
def _backend_ingest_url(endpoint: str) -> str:
|
|
33
|
+
if endpoint.endswith("/ingest/backend"):
|
|
34
|
+
return endpoint
|
|
35
|
+
if endpoint.endswith("/api/v1"):
|
|
36
|
+
return f"{endpoint}/ingest/backend"
|
|
37
|
+
return f"{endpoint}/api/v1/ingest/backend"
|
|
38
|
+
|
|
39
|
+
|
|
40
|
+
class CluePythonClient:
|
|
41
|
+
def __init__(
|
|
42
|
+
self,
|
|
43
|
+
*,
|
|
44
|
+
endpoint: str,
|
|
45
|
+
project_key: str,
|
|
46
|
+
environment: str,
|
|
47
|
+
producer_id: str = "python-service",
|
|
48
|
+
api_key: str | None = None,
|
|
49
|
+
timeout_seconds: float = 5.0,
|
|
50
|
+
max_event_bytes: int | None = DEFAULT_MAX_EVENT_BYTES,
|
|
51
|
+
max_payload_bytes: int = DEFAULT_MAX_PAYLOAD_BYTES,
|
|
52
|
+
flush_threshold_bytes: int = DEFAULT_FLUSH_THRESHOLD_BYTES,
|
|
53
|
+
max_buffer_bytes: int = DEFAULT_MAX_BUFFER_BYTES,
|
|
54
|
+
max_buffer_events: int = DEFAULT_MAX_BUFFER_EVENTS,
|
|
55
|
+
flush_interval_ms: int = DEFAULT_FLUSH_INTERVAL_MS,
|
|
56
|
+
auto_flush: bool = False,
|
|
57
|
+
opener: Callable[..., object] | None = None,
|
|
58
|
+
) -> None:
|
|
59
|
+
self._endpoint = endpoint.rstrip("/")
|
|
60
|
+
self._project_key = project_key
|
|
61
|
+
self._environment = environment
|
|
62
|
+
self._producer_id = producer_id
|
|
63
|
+
self._api_key = api_key
|
|
64
|
+
self._timeout_seconds = timeout_seconds
|
|
65
|
+
self._max_event_bytes = max_event_bytes
|
|
66
|
+
self._max_payload_bytes = max(1024, max_payload_bytes)
|
|
67
|
+
self._flush_threshold_bytes = max(1024, min(flush_threshold_bytes, self._max_payload_bytes))
|
|
68
|
+
self._max_buffer_bytes = max(1024, max_buffer_bytes)
|
|
69
|
+
self._max_buffer_events = max(1, max_buffer_events)
|
|
70
|
+
self._flush_interval_ms = max(1, flush_interval_ms)
|
|
71
|
+
self._auto_flush = auto_flush
|
|
72
|
+
self._opener = opener or urlopen
|
|
73
|
+
self._buffer: list[Mapping[str, JsonValue]] = []
|
|
74
|
+
self._buffer_bytes = 0
|
|
75
|
+
self._failed_batches: list[tuple[_BatchMetadata, list[Mapping[str, JsonValue]]]] = []
|
|
76
|
+
self._idle_timer: threading.Timer | None = None
|
|
77
|
+
self._lock = threading.RLock()
|
|
78
|
+
self._dropped_oversized_count = 0
|
|
79
|
+
self._parameter_snapshot_oversized_dropped = 0
|
|
80
|
+
self._dropped_buffer_full_count = 0
|
|
81
|
+
self._events_summarized_count = 0
|
|
82
|
+
self._sampling_dropped_count = 0
|
|
83
|
+
self._flush_failed_count = 0
|
|
84
|
+
self._flush_retry_count = 0
|
|
85
|
+
self._last_flush_status = "skipped"
|
|
86
|
+
|
|
87
|
+
def add_event(self, event: Mapping[str, JsonValue]) -> bool:
|
|
88
|
+
normalized_event = normalize_event_for_size(event, self._max_event_bytes)
|
|
89
|
+
if normalized_event is None:
|
|
90
|
+
self._dropped_oversized_count += 1
|
|
91
|
+
if event.get("event_category") in {
|
|
92
|
+
EVENT_CATEGORY_REQUEST,
|
|
93
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST,
|
|
94
|
+
}:
|
|
95
|
+
self._parameter_snapshot_oversized_dropped += 1
|
|
96
|
+
return False
|
|
97
|
+
|
|
98
|
+
event_bytes = self._measure_event_bytes(normalized_event)
|
|
99
|
+
with self._lock:
|
|
100
|
+
while self._buffer and (
|
|
101
|
+
len(self._buffer) >= self._max_buffer_events
|
|
102
|
+
or self._buffer_bytes + event_bytes > self._max_buffer_bytes
|
|
103
|
+
):
|
|
104
|
+
dropped = self._buffer.pop(0)
|
|
105
|
+
self._buffer_bytes = max(
|
|
106
|
+
0,
|
|
107
|
+
self._buffer_bytes - self._measure_event_bytes(dropped),
|
|
108
|
+
)
|
|
109
|
+
self._dropped_buffer_full_count += 1
|
|
110
|
+
|
|
111
|
+
if (
|
|
112
|
+
len(self._buffer) >= self._max_buffer_events
|
|
113
|
+
or self._buffer_bytes + event_bytes > self._max_buffer_bytes
|
|
114
|
+
):
|
|
115
|
+
self._dropped_buffer_full_count += 1
|
|
116
|
+
return False
|
|
117
|
+
|
|
118
|
+
self._buffer.append(normalized_event)
|
|
119
|
+
self._buffer_bytes += event_bytes
|
|
120
|
+
|
|
121
|
+
if self._auto_flush and self._buffer_bytes >= self._flush_threshold_bytes:
|
|
122
|
+
self._cancel_idle_timer()
|
|
123
|
+
should_flush = True
|
|
124
|
+
else:
|
|
125
|
+
self._schedule_idle_flush_locked()
|
|
126
|
+
should_flush = False
|
|
127
|
+
|
|
128
|
+
if should_flush:
|
|
129
|
+
try:
|
|
130
|
+
self.flush()
|
|
131
|
+
except Exception:
|
|
132
|
+
return True
|
|
133
|
+
return True
|
|
134
|
+
|
|
135
|
+
def record_events_summarized(self, count: int = 1) -> None:
|
|
136
|
+
self._events_summarized_count += max(0, int(count))
|
|
137
|
+
|
|
138
|
+
def record_sampling_dropped(self, count: int = 1) -> None:
|
|
139
|
+
self._sampling_dropped_count += max(0, int(count))
|
|
140
|
+
|
|
141
|
+
def build_envelope(
|
|
142
|
+
self,
|
|
143
|
+
events: Sequence[Mapping[str, JsonValue]] | None = None,
|
|
144
|
+
metadata: _BatchMetadata | None = None,
|
|
145
|
+
) -> dict[str, JsonValue]:
|
|
146
|
+
resolved_metadata = metadata or self._create_batch_metadata()
|
|
147
|
+
return build_backend_envelope(
|
|
148
|
+
project_key=self._project_key,
|
|
149
|
+
environment=self._environment,
|
|
150
|
+
producer_id=self._producer_id,
|
|
151
|
+
api_key=self._api_key,
|
|
152
|
+
batch_id=resolved_metadata.batch_id,
|
|
153
|
+
idempotency_key=resolved_metadata.idempotency_key,
|
|
154
|
+
sent_at=resolved_metadata.sent_at,
|
|
155
|
+
events=events if events is not None else self.pending_events(),
|
|
156
|
+
)
|
|
157
|
+
|
|
158
|
+
def flush(self, timeout_seconds: float | None = None) -> dict[str, JsonValue]:
|
|
159
|
+
self._cancel_idle_timer()
|
|
160
|
+
result: dict[str, JsonValue] = {}
|
|
161
|
+
while True:
|
|
162
|
+
batch = self._next_batch()
|
|
163
|
+
if batch is None:
|
|
164
|
+
return result
|
|
165
|
+
metadata, events = batch
|
|
166
|
+
envelope = self.build_envelope(events=events, metadata=metadata)
|
|
167
|
+
result = self._send_envelope(
|
|
168
|
+
envelope,
|
|
169
|
+
timeout_seconds=timeout_seconds or self._timeout_seconds,
|
|
170
|
+
on_failure=lambda: self._restore_failed_batch(metadata, events),
|
|
171
|
+
)
|
|
172
|
+
|
|
173
|
+
def _send_envelope(
|
|
174
|
+
self,
|
|
175
|
+
envelope: Mapping[str, JsonValue],
|
|
176
|
+
*,
|
|
177
|
+
timeout_seconds: float,
|
|
178
|
+
on_failure: Callable[[], None],
|
|
179
|
+
) -> dict[str, JsonValue]:
|
|
180
|
+
request = Request(
|
|
181
|
+
url=_backend_ingest_url(self._endpoint),
|
|
182
|
+
data=_json_bytes(envelope),
|
|
183
|
+
headers={
|
|
184
|
+
"content-type": "application/json",
|
|
185
|
+
"x-clue-project-key": self._project_key,
|
|
186
|
+
**({"x-clue-api-key": self._api_key} if self._api_key else {}),
|
|
187
|
+
},
|
|
188
|
+
method="POST",
|
|
189
|
+
)
|
|
190
|
+
|
|
191
|
+
try:
|
|
192
|
+
response = self._opener(request, timeout=timeout_seconds)
|
|
193
|
+
body = response.read()
|
|
194
|
+
except Exception:
|
|
195
|
+
self._flush_failed_count += 1
|
|
196
|
+
self._last_flush_status = "failed"
|
|
197
|
+
on_failure()
|
|
198
|
+
raise
|
|
199
|
+
self._last_flush_status = "success"
|
|
200
|
+
if not body:
|
|
201
|
+
return {}
|
|
202
|
+
return json.loads(body.decode("utf-8"))
|
|
203
|
+
|
|
204
|
+
def pending_events(self) -> Sequence[Mapping[str, JsonValue]]:
|
|
205
|
+
with self._lock:
|
|
206
|
+
failed_events = [
|
|
207
|
+
event
|
|
208
|
+
for _metadata, events in self._failed_batches
|
|
209
|
+
for event in events
|
|
210
|
+
]
|
|
211
|
+
return tuple([*failed_events, *self._buffer])
|
|
212
|
+
|
|
213
|
+
def drain_events(self) -> Sequence[Mapping[str, JsonValue]]:
|
|
214
|
+
with self._lock:
|
|
215
|
+
events = tuple(self._buffer)
|
|
216
|
+
self._buffer.clear()
|
|
217
|
+
self._buffer_bytes = 0
|
|
218
|
+
self._cancel_idle_timer()
|
|
219
|
+
return events
|
|
220
|
+
|
|
221
|
+
def close(self) -> None:
|
|
222
|
+
self._cancel_idle_timer()
|
|
223
|
+
|
|
224
|
+
def internal_metrics(self) -> dict[str, JsonValue]:
|
|
225
|
+
queue_size = len(self.pending_events())
|
|
226
|
+
return {
|
|
227
|
+
"events_dropped_count": (
|
|
228
|
+
self._dropped_oversized_count + self._dropped_buffer_full_count
|
|
229
|
+
),
|
|
230
|
+
"events_summarized_count": self._events_summarized_count,
|
|
231
|
+
"sampling_dropped_count": self._sampling_dropped_count,
|
|
232
|
+
"oversized_event_count": self._dropped_oversized_count,
|
|
233
|
+
"snapshot_truncated_count": self._parameter_snapshot_oversized_dropped,
|
|
234
|
+
"flush_failed_count": self._flush_failed_count,
|
|
235
|
+
"flush_retry_count": self._flush_retry_count,
|
|
236
|
+
"queue_size": queue_size,
|
|
237
|
+
"queue_bytes": self._buffer_bytes,
|
|
238
|
+
"max_payload_bytes": self._max_payload_bytes,
|
|
239
|
+
"flush_threshold_bytes": self._flush_threshold_bytes,
|
|
240
|
+
"last_flush_status": self._last_flush_status,
|
|
241
|
+
"dropped_oversized_count": self._dropped_oversized_count,
|
|
242
|
+
"parameter_snapshot_oversized_dropped": self._parameter_snapshot_oversized_dropped,
|
|
243
|
+
"dropped_buffer_full_count": self._dropped_buffer_full_count,
|
|
244
|
+
"flush_interval_ms": self._flush_interval_ms,
|
|
245
|
+
}
|
|
246
|
+
|
|
247
|
+
def _create_batch_metadata(self) -> _BatchMetadata:
|
|
248
|
+
batch_id = str(uuid.uuid4())
|
|
249
|
+
return _BatchMetadata(
|
|
250
|
+
batch_id=f"batch_{batch_id}",
|
|
251
|
+
idempotency_key=f"idem_{batch_id}",
|
|
252
|
+
sent_at=self._now_iso(),
|
|
253
|
+
)
|
|
254
|
+
|
|
255
|
+
def _next_batch(self) -> tuple[_BatchMetadata, list[Mapping[str, JsonValue]]] | None:
|
|
256
|
+
with self._lock:
|
|
257
|
+
if self._failed_batches:
|
|
258
|
+
return self._failed_batches.pop(0)
|
|
259
|
+
if not self._buffer:
|
|
260
|
+
return None
|
|
261
|
+
|
|
262
|
+
metadata = self._create_batch_metadata()
|
|
263
|
+
batch: list[Mapping[str, JsonValue]] = []
|
|
264
|
+
while self._buffer:
|
|
265
|
+
candidate = [*batch, self._buffer[0]]
|
|
266
|
+
if batch and self._measure_envelope_bytes(candidate, metadata) > self._max_payload_bytes:
|
|
267
|
+
break
|
|
268
|
+
event = self._buffer.pop(0)
|
|
269
|
+
self._buffer_bytes = max(
|
|
270
|
+
0,
|
|
271
|
+
self._buffer_bytes - self._measure_event_bytes(event),
|
|
272
|
+
)
|
|
273
|
+
batch.append(event)
|
|
274
|
+
|
|
275
|
+
if not batch:
|
|
276
|
+
event = self._buffer.pop(0)
|
|
277
|
+
self._buffer_bytes = max(
|
|
278
|
+
0,
|
|
279
|
+
self._buffer_bytes - self._measure_event_bytes(event),
|
|
280
|
+
)
|
|
281
|
+
self._dropped_oversized_count += 1
|
|
282
|
+
return None
|
|
283
|
+
return metadata, batch
|
|
284
|
+
|
|
285
|
+
def _restore_failed_batch(
|
|
286
|
+
self,
|
|
287
|
+
metadata: _BatchMetadata,
|
|
288
|
+
events: list[Mapping[str, JsonValue]],
|
|
289
|
+
) -> None:
|
|
290
|
+
with self._lock:
|
|
291
|
+
self._failed_batches.insert(0, (metadata, events))
|
|
292
|
+
if self._auto_flush:
|
|
293
|
+
self._schedule_idle_flush_locked()
|
|
294
|
+
|
|
295
|
+
def _schedule_idle_flush_locked(self) -> None:
|
|
296
|
+
if not self._auto_flush:
|
|
297
|
+
return
|
|
298
|
+
if self._idle_timer is not None:
|
|
299
|
+
self._idle_timer.cancel()
|
|
300
|
+
self._idle_timer = threading.Timer(
|
|
301
|
+
self._flush_interval_ms / 1000,
|
|
302
|
+
self._flush_from_timer,
|
|
303
|
+
)
|
|
304
|
+
self._idle_timer.daemon = True
|
|
305
|
+
self._idle_timer.start()
|
|
306
|
+
|
|
307
|
+
def _cancel_idle_timer(self) -> None:
|
|
308
|
+
if self._idle_timer is not None:
|
|
309
|
+
self._idle_timer.cancel()
|
|
310
|
+
self._idle_timer = None
|
|
311
|
+
|
|
312
|
+
def _flush_from_timer(self) -> None:
|
|
313
|
+
with self._lock:
|
|
314
|
+
self._idle_timer = None
|
|
315
|
+
try:
|
|
316
|
+
self.flush()
|
|
317
|
+
except Exception:
|
|
318
|
+
return
|
|
319
|
+
|
|
320
|
+
def _measure_event_bytes(self, event: Mapping[str, JsonValue]) -> int:
|
|
321
|
+
return len(_json_bytes(event))
|
|
322
|
+
|
|
323
|
+
def _measure_envelope_bytes(
|
|
324
|
+
self,
|
|
325
|
+
events: Sequence[Mapping[str, JsonValue]],
|
|
326
|
+
metadata: _BatchMetadata,
|
|
327
|
+
) -> int:
|
|
328
|
+
return len(
|
|
329
|
+
_json_bytes(self.build_envelope(events=events, metadata=metadata))
|
|
330
|
+
)
|
|
331
|
+
|
|
332
|
+
def _now_iso(self) -> str:
|
|
333
|
+
from datetime import datetime, timezone
|
|
334
|
+
|
|
335
|
+
return datetime.now(timezone.utc).isoformat().replace("+00:00", "Z")
|
|
@@ -0,0 +1,224 @@
|
|
|
1
|
+
from __future__ import annotations
|
|
2
|
+
|
|
3
|
+
BACKEND_STATUS_OBSERVED = "observed"
|
|
4
|
+
BACKEND_STATUS_STARTED = "started"
|
|
5
|
+
BACKEND_STATUS_FINISHED = "finished"
|
|
6
|
+
BACKEND_STATUS_FAILED = "failed"
|
|
7
|
+
BACKEND_STATUSES = (
|
|
8
|
+
BACKEND_STATUS_OBSERVED,
|
|
9
|
+
BACKEND_STATUS_STARTED,
|
|
10
|
+
BACKEND_STATUS_FINISHED,
|
|
11
|
+
BACKEND_STATUS_FAILED,
|
|
12
|
+
)
|
|
13
|
+
|
|
14
|
+
BACKEND_SDK_TYPE = "backend"
|
|
15
|
+
BACKEND_SDK_VERSION = "0.0.1"
|
|
16
|
+
BACKEND_SDK_SCHEMA_VERSION = 1
|
|
17
|
+
BACKEND_RUNTIME_LANGUAGE = "python"
|
|
18
|
+
|
|
19
|
+
SDK_COLLECTION_MODE_STANDARD = "standard"
|
|
20
|
+
SDK_COLLECTION_MODE_DIAGNOSTIC = "diagnostic"
|
|
21
|
+
SDK_COLLECTION_MODES = (
|
|
22
|
+
SDK_COLLECTION_MODE_STANDARD,
|
|
23
|
+
SDK_COLLECTION_MODE_DIAGNOSTIC,
|
|
24
|
+
)
|
|
25
|
+
|
|
26
|
+
EVENT_CATEGORY_REQUEST = "request"
|
|
27
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST = "outbound_request"
|
|
28
|
+
EVENT_CATEGORY_ERROR = "error"
|
|
29
|
+
EVENT_CATEGORY_JOB = "job"
|
|
30
|
+
EVENT_CATEGORY_TOOL_CALL = "tool_call"
|
|
31
|
+
EVENT_CATEGORY_DOMAIN_COMMAND = "domain_command"
|
|
32
|
+
EVENT_CATEGORY_REPOSITORY_MUTATION = "repository_mutation"
|
|
33
|
+
EVENT_CATEGORY_STATE_TRANSITION = "state_transition"
|
|
34
|
+
EVENT_CATEGORY_CUSTOM = "custom"
|
|
35
|
+
BACKEND_EVENT_CATEGORIES = (
|
|
36
|
+
EVENT_CATEGORY_REQUEST,
|
|
37
|
+
EVENT_CATEGORY_ERROR,
|
|
38
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST,
|
|
39
|
+
EVENT_CATEGORY_JOB,
|
|
40
|
+
EVENT_CATEGORY_TOOL_CALL,
|
|
41
|
+
EVENT_CATEGORY_DOMAIN_COMMAND,
|
|
42
|
+
EVENT_CATEGORY_REPOSITORY_MUTATION,
|
|
43
|
+
EVENT_CATEGORY_STATE_TRANSITION,
|
|
44
|
+
EVENT_CATEGORY_CUSTOM,
|
|
45
|
+
)
|
|
46
|
+
PYTHON_RUNTIME_EVENT_CATEGORIES = (
|
|
47
|
+
EVENT_CATEGORY_REQUEST,
|
|
48
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST,
|
|
49
|
+
EVENT_CATEGORY_ERROR,
|
|
50
|
+
EVENT_CATEGORY_JOB,
|
|
51
|
+
EVENT_CATEGORY_TOOL_CALL,
|
|
52
|
+
EVENT_CATEGORY_DOMAIN_COMMAND,
|
|
53
|
+
EVENT_CATEGORY_REPOSITORY_MUTATION,
|
|
54
|
+
EVENT_CATEGORY_STATE_TRANSITION,
|
|
55
|
+
EVENT_CATEGORY_CUSTOM,
|
|
56
|
+
)
|
|
57
|
+
|
|
58
|
+
REQUEST_EVENT_STARTED = "request_started"
|
|
59
|
+
REQUEST_EVENT_FINISHED = "request_finished"
|
|
60
|
+
REQUEST_EVENT_FAILED = "request_failed"
|
|
61
|
+
|
|
62
|
+
OUTBOUND_REQUEST_EVENT_STARTED = "outbound_request_started"
|
|
63
|
+
OUTBOUND_REQUEST_EVENT_FINISHED = "outbound_request_finished"
|
|
64
|
+
OUTBOUND_REQUEST_EVENT_FAILED = "outbound_request_failed"
|
|
65
|
+
|
|
66
|
+
JOB_EVENT_STARTED = "job_started"
|
|
67
|
+
JOB_EVENT_FINISHED = "job_finished"
|
|
68
|
+
JOB_EVENT_FAILED = "job_failed"
|
|
69
|
+
|
|
70
|
+
TOOL_CALL_EVENT_STARTED = "tool_call_started"
|
|
71
|
+
TOOL_CALL_EVENT_FINISHED = "tool_call_finished"
|
|
72
|
+
TOOL_CALL_EVENT_FAILED = "tool_call_failed"
|
|
73
|
+
|
|
74
|
+
DOMAIN_COMMAND_EVENT_STARTED = "domain_command_started"
|
|
75
|
+
DOMAIN_COMMAND_EVENT_FINISHED = "domain_command_finished"
|
|
76
|
+
DOMAIN_COMMAND_EVENT_FAILED = "domain_command_failed"
|
|
77
|
+
|
|
78
|
+
REPOSITORY_MUTATION_EVENT_STARTED = "repository_mutation_started"
|
|
79
|
+
REPOSITORY_MUTATION_EVENT_FINISHED = "repository_mutation_finished"
|
|
80
|
+
REPOSITORY_MUTATION_EVENT_FAILED = "repository_mutation_failed"
|
|
81
|
+
|
|
82
|
+
STATE_TRANSITION_EVENT_STARTED = "state_transition_started"
|
|
83
|
+
STATE_TRANSITION_EVENT_FINISHED = "state_transition_finished"
|
|
84
|
+
STATE_TRANSITION_EVENT_FAILED = "state_transition_failed"
|
|
85
|
+
|
|
86
|
+
BACKEND_ERROR_EVENT = "backend_error"
|
|
87
|
+
CUSTOM_EVENT_EMITTED = "custom_emitted"
|
|
88
|
+
REPOSITORY_MUTATION_SUMMARY_EVENT = "repository_mutation_summary"
|
|
89
|
+
QUEUE_LIFECYCLE_SUMMARY_EVENT = "queue_lifecycle_summary"
|
|
90
|
+
DEPENDENCY_SUMMARY_EVENT = "dependency_summary"
|
|
91
|
+
CACHE_OPERATION_SUMMARY_EVENT = "cache_operation_summary"
|
|
92
|
+
APPLICATION_LOG_SUMMARY_EVENT = "application_log_summary"
|
|
93
|
+
SDK_DIAGNOSTIC_EVENT = "sdk_diagnostic_observed"
|
|
94
|
+
CUSTOM_EVENT_SDK_INITIALIZED = "sdk_initialized"
|
|
95
|
+
CUSTOM_EVENT_IDENTITY_IDENTIFIED = "identity_identified"
|
|
96
|
+
CUSTOM_EVENT_ACCOUNT_ASSOCIATED = "account_associated"
|
|
97
|
+
CUSTOM_EVENT_IDENTITY_LOGGED_OUT = "identity_logged_out"
|
|
98
|
+
BACKEND_OFFICIAL_CUSTOM_EVENT_NAMES = (
|
|
99
|
+
CUSTOM_EVENT_SDK_INITIALIZED,
|
|
100
|
+
CUSTOM_EVENT_IDENTITY_IDENTIFIED,
|
|
101
|
+
CUSTOM_EVENT_ACCOUNT_ASSOCIATED,
|
|
102
|
+
CUSTOM_EVENT_IDENTITY_LOGGED_OUT,
|
|
103
|
+
)
|
|
104
|
+
|
|
105
|
+
JOB_EVENT_BY_STATUS = {
|
|
106
|
+
BACKEND_STATUS_STARTED: JOB_EVENT_STARTED,
|
|
107
|
+
BACKEND_STATUS_FINISHED: JOB_EVENT_FINISHED,
|
|
108
|
+
BACKEND_STATUS_FAILED: JOB_EVENT_FAILED,
|
|
109
|
+
}
|
|
110
|
+
|
|
111
|
+
TOOL_CALL_EVENT_BY_STATUS = {
|
|
112
|
+
BACKEND_STATUS_STARTED: TOOL_CALL_EVENT_STARTED,
|
|
113
|
+
BACKEND_STATUS_FINISHED: TOOL_CALL_EVENT_FINISHED,
|
|
114
|
+
BACKEND_STATUS_FAILED: TOOL_CALL_EVENT_FAILED,
|
|
115
|
+
}
|
|
116
|
+
|
|
117
|
+
DOMAIN_COMMAND_EVENT_BY_STATUS = {
|
|
118
|
+
BACKEND_STATUS_STARTED: DOMAIN_COMMAND_EVENT_STARTED,
|
|
119
|
+
BACKEND_STATUS_FINISHED: DOMAIN_COMMAND_EVENT_FINISHED,
|
|
120
|
+
BACKEND_STATUS_FAILED: DOMAIN_COMMAND_EVENT_FAILED,
|
|
121
|
+
}
|
|
122
|
+
|
|
123
|
+
REPOSITORY_MUTATION_EVENT_BY_STATUS = {
|
|
124
|
+
BACKEND_STATUS_STARTED: REPOSITORY_MUTATION_EVENT_STARTED,
|
|
125
|
+
BACKEND_STATUS_FINISHED: REPOSITORY_MUTATION_EVENT_FINISHED,
|
|
126
|
+
BACKEND_STATUS_FAILED: REPOSITORY_MUTATION_EVENT_FAILED,
|
|
127
|
+
}
|
|
128
|
+
|
|
129
|
+
STATE_TRANSITION_EVENT_BY_STATUS = {
|
|
130
|
+
BACKEND_STATUS_STARTED: STATE_TRANSITION_EVENT_STARTED,
|
|
131
|
+
BACKEND_STATUS_FINISHED: STATE_TRANSITION_EVENT_FINISHED,
|
|
132
|
+
BACKEND_STATUS_FAILED: STATE_TRANSITION_EVENT_FAILED,
|
|
133
|
+
}
|
|
134
|
+
|
|
135
|
+
PYTHON_RUNTIME_EVENT_NAMES = (
|
|
136
|
+
REQUEST_EVENT_STARTED,
|
|
137
|
+
REQUEST_EVENT_FINISHED,
|
|
138
|
+
REQUEST_EVENT_FAILED,
|
|
139
|
+
OUTBOUND_REQUEST_EVENT_STARTED,
|
|
140
|
+
OUTBOUND_REQUEST_EVENT_FINISHED,
|
|
141
|
+
OUTBOUND_REQUEST_EVENT_FAILED,
|
|
142
|
+
JOB_EVENT_STARTED,
|
|
143
|
+
JOB_EVENT_FINISHED,
|
|
144
|
+
JOB_EVENT_FAILED,
|
|
145
|
+
TOOL_CALL_EVENT_STARTED,
|
|
146
|
+
TOOL_CALL_EVENT_FINISHED,
|
|
147
|
+
TOOL_CALL_EVENT_FAILED,
|
|
148
|
+
DOMAIN_COMMAND_EVENT_STARTED,
|
|
149
|
+
DOMAIN_COMMAND_EVENT_FINISHED,
|
|
150
|
+
DOMAIN_COMMAND_EVENT_FAILED,
|
|
151
|
+
REPOSITORY_MUTATION_EVENT_STARTED,
|
|
152
|
+
REPOSITORY_MUTATION_EVENT_FINISHED,
|
|
153
|
+
REPOSITORY_MUTATION_EVENT_FAILED,
|
|
154
|
+
STATE_TRANSITION_EVENT_STARTED,
|
|
155
|
+
STATE_TRANSITION_EVENT_FINISHED,
|
|
156
|
+
STATE_TRANSITION_EVENT_FAILED,
|
|
157
|
+
BACKEND_ERROR_EVENT,
|
|
158
|
+
CUSTOM_EVENT_EMITTED,
|
|
159
|
+
)
|
|
160
|
+
BACKEND_EVENT_NAMES = (
|
|
161
|
+
REQUEST_EVENT_STARTED,
|
|
162
|
+
REQUEST_EVENT_FINISHED,
|
|
163
|
+
REQUEST_EVENT_FAILED,
|
|
164
|
+
OUTBOUND_REQUEST_EVENT_STARTED,
|
|
165
|
+
OUTBOUND_REQUEST_EVENT_FINISHED,
|
|
166
|
+
OUTBOUND_REQUEST_EVENT_FAILED,
|
|
167
|
+
JOB_EVENT_STARTED,
|
|
168
|
+
JOB_EVENT_FINISHED,
|
|
169
|
+
JOB_EVENT_FAILED,
|
|
170
|
+
TOOL_CALL_EVENT_STARTED,
|
|
171
|
+
TOOL_CALL_EVENT_FINISHED,
|
|
172
|
+
TOOL_CALL_EVENT_FAILED,
|
|
173
|
+
DOMAIN_COMMAND_EVENT_STARTED,
|
|
174
|
+
DOMAIN_COMMAND_EVENT_FINISHED,
|
|
175
|
+
DOMAIN_COMMAND_EVENT_FAILED,
|
|
176
|
+
REPOSITORY_MUTATION_EVENT_STARTED,
|
|
177
|
+
REPOSITORY_MUTATION_EVENT_FINISHED,
|
|
178
|
+
REPOSITORY_MUTATION_EVENT_FAILED,
|
|
179
|
+
STATE_TRANSITION_EVENT_STARTED,
|
|
180
|
+
STATE_TRANSITION_EVENT_FINISHED,
|
|
181
|
+
STATE_TRANSITION_EVENT_FAILED,
|
|
182
|
+
BACKEND_ERROR_EVENT,
|
|
183
|
+
REPOSITORY_MUTATION_SUMMARY_EVENT,
|
|
184
|
+
QUEUE_LIFECYCLE_SUMMARY_EVENT,
|
|
185
|
+
DEPENDENCY_SUMMARY_EVENT,
|
|
186
|
+
CACHE_OPERATION_SUMMARY_EVENT,
|
|
187
|
+
APPLICATION_LOG_SUMMARY_EVENT,
|
|
188
|
+
SDK_DIAGNOSTIC_EVENT,
|
|
189
|
+
CUSTOM_EVENT_EMITTED,
|
|
190
|
+
)
|
|
191
|
+
|
|
192
|
+
SURFACE_TYPE_API = "api"
|
|
193
|
+
SURFACE_TYPE_SERVICE = "service"
|
|
194
|
+
SURFACE_TYPE_JOB = "job"
|
|
195
|
+
BACKEND_SURFACE_TYPES = (
|
|
196
|
+
SURFACE_TYPE_API,
|
|
197
|
+
SURFACE_TYPE_SERVICE,
|
|
198
|
+
SURFACE_TYPE_JOB,
|
|
199
|
+
)
|
|
200
|
+
|
|
201
|
+
CAPTURE_MODE_ALLOWED_PLAINTEXT = "allowed_plaintext"
|
|
202
|
+
CAPTURE_MODE_MASKED_FINGERPRINT = "masked_fingerprint"
|
|
203
|
+
CAPTURE_MODE_FORBIDDEN = "forbidden"
|
|
204
|
+
CAPTURE_MODE_TRUNCATED = "truncated"
|
|
205
|
+
BACKEND_CAPTURE_MODES = (
|
|
206
|
+
CAPTURE_MODE_ALLOWED_PLAINTEXT,
|
|
207
|
+
CAPTURE_MODE_MASKED_FINGERPRINT,
|
|
208
|
+
CAPTURE_MODE_FORBIDDEN,
|
|
209
|
+
CAPTURE_MODE_TRUNCATED,
|
|
210
|
+
)
|
|
211
|
+
|
|
212
|
+
BACKEND_LIFECYCLE_MARKERS = (
|
|
213
|
+
"start",
|
|
214
|
+
"finish",
|
|
215
|
+
"fail",
|
|
216
|
+
)
|
|
217
|
+
|
|
218
|
+
BACKEND_MUTATION_KINDS = (
|
|
219
|
+
"create",
|
|
220
|
+
"update",
|
|
221
|
+
"delete",
|
|
222
|
+
"upsert",
|
|
223
|
+
"bulk_write",
|
|
224
|
+
)
|
|
@@ -0,0 +1,96 @@
|
|
|
1
|
+
from __future__ import annotations
|
|
2
|
+
|
|
3
|
+
import json
|
|
4
|
+
from collections.abc import Mapping
|
|
5
|
+
|
|
6
|
+
from .contracts import (
|
|
7
|
+
CAPTURE_MODE_TRUNCATED,
|
|
8
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST,
|
|
9
|
+
EVENT_CATEGORY_REQUEST,
|
|
10
|
+
)
|
|
11
|
+
from .privacy import JsonValue
|
|
12
|
+
|
|
13
|
+
DEFAULT_MAX_EVENT_BYTES = 128 * 1024
|
|
14
|
+
TRUNCATED_PARAMETER_VALUE = "__truncated__"
|
|
15
|
+
PARAMETER_SNAPSHOT_FIELDS = (
|
|
16
|
+
"request_parameters_snapshot_json",
|
|
17
|
+
"response_parameters_snapshot_json",
|
|
18
|
+
)
|
|
19
|
+
|
|
20
|
+
|
|
21
|
+
def measure_event_bytes(event: Mapping[str, JsonValue]) -> int:
|
|
22
|
+
return len(
|
|
23
|
+
json.dumps(event, ensure_ascii=False, separators=(",", ":")).encode("utf-8")
|
|
24
|
+
)
|
|
25
|
+
|
|
26
|
+
|
|
27
|
+
def _is_parameter_leaf(value: JsonValue) -> bool:
|
|
28
|
+
return (
|
|
29
|
+
isinstance(value, dict)
|
|
30
|
+
and isinstance(value.get("type"), str)
|
|
31
|
+
and isinstance(value.get("capture_mode"), str)
|
|
32
|
+
and "value" in value
|
|
33
|
+
)
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
def _truncate_parameter_snapshot(value: JsonValue) -> JsonValue:
|
|
37
|
+
if isinstance(value, list):
|
|
38
|
+
return [_truncate_parameter_snapshot(entry) for entry in value]
|
|
39
|
+
|
|
40
|
+
if _is_parameter_leaf(value):
|
|
41
|
+
return {
|
|
42
|
+
"type": value.get("type"),
|
|
43
|
+
"value": TRUNCATED_PARAMETER_VALUE,
|
|
44
|
+
"capture_mode": CAPTURE_MODE_TRUNCATED,
|
|
45
|
+
}
|
|
46
|
+
|
|
47
|
+
if isinstance(value, dict):
|
|
48
|
+
return {
|
|
49
|
+
str(key): _truncate_parameter_snapshot(child)
|
|
50
|
+
for key, child in value.items()
|
|
51
|
+
}
|
|
52
|
+
|
|
53
|
+
return value
|
|
54
|
+
|
|
55
|
+
|
|
56
|
+
def _is_parameter_snapshot_event(event: Mapping[str, JsonValue]) -> bool:
|
|
57
|
+
return event.get("event_category") in {
|
|
58
|
+
EVENT_CATEGORY_REQUEST,
|
|
59
|
+
EVENT_CATEGORY_OUTBOUND_REQUEST,
|
|
60
|
+
}
|
|
61
|
+
|
|
62
|
+
|
|
63
|
+
def normalize_event_for_size(
|
|
64
|
+
event: Mapping[str, JsonValue],
|
|
65
|
+
max_event_bytes: int | None,
|
|
66
|
+
) -> dict[str, JsonValue] | None:
|
|
67
|
+
normalized = dict(event)
|
|
68
|
+
if max_event_bytes is None or max_event_bytes <= 0:
|
|
69
|
+
return normalized
|
|
70
|
+
|
|
71
|
+
if measure_event_bytes(normalized) <= max_event_bytes:
|
|
72
|
+
return normalized
|
|
73
|
+
|
|
74
|
+
if not _is_parameter_snapshot_event(normalized):
|
|
75
|
+
return None
|
|
76
|
+
|
|
77
|
+
has_snapshot = any(field in normalized for field in PARAMETER_SNAPSHOT_FIELDS)
|
|
78
|
+
if not has_snapshot:
|
|
79
|
+
return None
|
|
80
|
+
|
|
81
|
+
truncated = dict(normalized)
|
|
82
|
+
for field in PARAMETER_SNAPSHOT_FIELDS:
|
|
83
|
+
value = truncated.get(field)
|
|
84
|
+
if value is not None:
|
|
85
|
+
truncated[field] = _truncate_parameter_snapshot(value)
|
|
86
|
+
|
|
87
|
+
if isinstance(truncated.get("properties"), dict):
|
|
88
|
+
properties = dict(truncated["properties"])
|
|
89
|
+
properties["degradation_stage"] = "parameter_snapshot_truncated"
|
|
90
|
+
truncated["properties"] = properties
|
|
91
|
+
else:
|
|
92
|
+
truncated["properties"] = {
|
|
93
|
+
"degradation_stage": "parameter_snapshot_truncated",
|
|
94
|
+
}
|
|
95
|
+
|
|
96
|
+
return truncated if measure_event_bytes(truncated) <= max_event_bytes else None
|