Repository navigation
Expand file tree
/
Copy pathupload.py
More file actions
474 lines (418 loc) · 18.8 KB
/
Copy pathupload.py
File metadata and controls
474 lines (418 loc) · 18.8 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
"""Upload router for the Memory Tracker API."""
from fastapi import APIRouter, Depends, HTTPException
from sqlalchemy.ext.asyncio import AsyncSession
from sqlalchemy import select, desc, func, update, and_, delete
from sqlalchemy.exc import IntegrityError
from datetime import datetime
import logging
from .. import schemas, crud, models
from ..database import get_database
from ..auth import get_current_token
router = APIRouter(prefix="/api", tags=["upload"])
async def cleanup_old_flamegraphs_if_needed(
db: AsyncSession, environment_id: str, max_flamegraphs: int = 100
):
"""
Set flamegraph_html to NULL for runs older than the last max_flamegraphs in this environment.
"""
logger = logging.getLogger(__name__)
keep_runs_result = await db.execute(
select(models.Run.run_id)
.where(models.Run.environment_id == environment_id)
.order_by(desc(models.Run.timestamp))
.limit(max_flamegraphs)
)
keep_run_ids = [row[0] for row in keep_runs_result.fetchall()]
if not keep_run_ids:
return 0
total_flamegraphs_result = await db.execute(
select(func.count())
.select_from(models.BenchmarkResult)
.join(models.Run)
.where(
models.Run.environment_id == environment_id,
models.BenchmarkResult.flamegraph_html.is_not(None),
)
)
total_flamegraphs_before = total_flamegraphs_result.scalar()
count_query = (
select(func.count())
.select_from(models.BenchmarkResult)
.where(
and_(
models.BenchmarkResult.run_id.not_in(keep_run_ids),
models.BenchmarkResult.flamegraph_html.is_not(None),
)
)
)
count_result = await db.execute(count_query)
rows_to_clean = count_result.scalar()
if rows_to_clean == 0:
logger.info(
f"No flamegraphs need cleaning for environment '{environment_id}' (all {total_flamegraphs_before} flamegraphs are from recent runs)"
)
return 0
cleanup_query = (
update(models.BenchmarkResult)
.where(
and_(
models.BenchmarkResult.run_id.not_in(keep_run_ids),
models.BenchmarkResult.flamegraph_html.is_not(None),
)
)
.values(flamegraph_html=None)
)
try:
await db.execute(cleanup_query)
await db.commit()
verify_result = await db.execute(count_query)
remaining = verify_result.scalar()
actual_cleaned = rows_to_clean - remaining
final_flamegraphs_result = await db.execute(
select(func.count())
.select_from(models.BenchmarkResult)
.join(models.Run)
.where(
models.Run.environment_id == environment_id,
models.BenchmarkResult.flamegraph_html.is_not(None),
)
)
total_flamegraphs_after = final_flamegraphs_result.scalar()
if actual_cleaned > 0:
logger.info(
f"Cleaned up {actual_cleaned} flamegraphs for environment '{environment_id}': {total_flamegraphs_before} → {total_flamegraphs_after} flamegraphs remaining"
)
else:
logger.error(
f"Cleanup FAILED for environment '{environment_id}'. Expected to clean {rows_to_clean}, but {remaining} still remain"
)
return actual_cleaned
except Exception as e:
logger.error(
f"Exception during cleanup for environment '{environment_id}': {e}"
)
await db.rollback()
raise
@router.post("/upload-run", response_model=dict)
async def upload_worker_run(
upload_data: schemas.WorkerRunUpload,
db: AsyncSession = Depends(get_database),
current_token: models.AuthToken = Depends(get_current_token),
):
logger = logging.getLogger(__name__)
logger.info(
f"Authenticated upload request from token '{current_token.name}' for binary_id='{upload_data.binary_id}', environment_id='{upload_data.environment_id}'"
)
logger.debug(
f"Upload contains {len(upload_data.benchmark_results)} benchmark results"
)
metadata = upload_data.metadata
# Extract commit information
commit_info = metadata.get("commit", {})
commit_sha = commit_info.get("hexsha")
if not commit_sha:
logger.error("Upload failed: Missing commit SHA in metadata")
raise HTTPException(status_code=400, detail="Missing commit SHA in metadata")
logger.info(f"Processing upload for commit {commit_sha[:8]}")
# Extract Python version
version_info = metadata.get("version", {})
python_version = schemas.PythonVersion(
major=version_info.get("major"),
minor=version_info.get("minor"),
patch=version_info.get(
"micro", 0
), # 'micro' in metadata maps to 'patch' in our schema
)
logger.debug(
f"Extracted Python version: {python_version.major}.{python_version.minor}.{python_version.patch}"
)
# Use provided binary_id and environment_id from worker
binary_id = upload_data.binary_id
environment_id = upload_data.environment_id
# Validate binary exists
logger.debug(f"Validating binary '{binary_id}' exists")
binary = await crud.get_binary_by_id(db, binary_id=binary_id)
if not binary:
logger.error(f"Upload failed: Binary '{binary_id}' not found")
raise HTTPException(
status_code=400,
detail=f"Binary '{binary_id}' not found. Binaries must be pre-registered.",
)
logger.info(f"Binary '{binary_id}' validated successfully")
# Validate environment exists
logger.debug(f"Validating environment '{environment_id}' exists")
environment = await crud.get_environment_by_id(db, environment_id=environment_id)
if not environment:
logger.error(f"Upload failed: Environment '{environment_id}' not found")
raise HTTPException(
status_code=400,
detail=f"Environment '{environment_id}' not found. Environments must be pre-registered.",
)
logger.info(f"Environment '{environment_id}' validated successfully")
# Validate configure flags - the registered binary flags must be a subset of uploaded flags
configure_vars = metadata.get("configure_vars", {})
uploaded_config_args = configure_vars.get("CONFIG_ARGS", "")
# Strip quotes from each flag and filter out non-configure flags
def clean_flag(flag):
"""Remove surrounding quotes from a flag."""
return flag.strip().strip("'\"")
# Split and clean uploaded flags, filtering out environment variables
raw_uploaded_flags = uploaded_config_args.split() if uploaded_config_args else []
uploaded_flags = set()
for flag in raw_uploaded_flags:
cleaned = clean_flag(flag)
# Only include flags that start with -- (configure flags)
if cleaned.startswith('--'):
uploaded_flags.add(cleaned)
# Clean registered flags too for consistency
registered_flags = {clean_flag(flag) for flag in (binary.flags or [])}
logger.debug(
f"Configure flags validation: registered={sorted(registered_flags)}, uploaded={sorted(uploaded_flags)}"
)
logger.debug(f"Raw CONFIG_ARGS from metadata: '{uploaded_config_args}'")
# Check if registered flags are a subset of uploaded flags
if registered_flags and not registered_flags.issubset(uploaded_flags):
missing_flags = registered_flags - uploaded_flags
logger.error(
f"Upload failed: Configure flags mismatch for binary '{binary_id}'. "
f"Missing flags: {sorted(missing_flags)}, "
f"Required: {sorted(registered_flags)}, "
f"Provided: {sorted(uploaded_flags)}, "
f"Raw CONFIG_ARGS: '{uploaded_config_args}'"
)
raise HTTPException(
status_code=400,
detail=f"Binary '{binary_id}' requires configure flags {sorted(list(registered_flags))} but upload only has {sorted(list(uploaded_flags))}. "
f"Registered configure flags must be a subset of upload configure flags.",
)
logger.info(f"Configure flags validation passed for binary '{binary_id}'")
# Note: Duplicate commits are now prevented by database unique constraint on (commit_sha, binary_id, environment_id)
# Create or get commit
logger.debug(f"Looking up commit {commit_sha[:8]} in database")
commit = await crud.get_commit_by_sha(db, sha=commit_sha)
if not commit:
logger.info(f"Commit {commit_sha[:8]} not found, creating new commit record")
# Create commit from metadata
commit_data = schemas.CommitCreate(
sha=commit_sha,
timestamp=datetime.fromisoformat(
commit_info.get("committed_date", "").replace("Z", "+00:00")
),
message=commit_info.get("message", ""),
author=commit_info.get("author", ""),
python_version=python_version,
)
try:
commit = await crud.create_commit(db, commit_data)
logger.info(f"Successfully created commit record for {commit_sha[:8]}")
except Exception as e:
logger.error(f"Failed to create commit record for {commit_sha[:8]}: {e}")
raise HTTPException(
status_code=500, detail=f"Failed to create commit record: {str(e)}"
)
else:
logger.debug(f"Found existing commit record for {commit_sha[:8]}")
# Create run
run_id = f"run_{commit_sha[:8]}_{binary_id}_{environment_id}_{int(datetime.now().timestamp())}"
logger.info(f"Creating run with ID: {run_id}")
run_data = schemas.RunCreate(
run_id=run_id,
commit_sha=commit_sha,
binary_id=binary_id,
environment_id=environment_id,
python_version=python_version,
timestamp=datetime.now(),
)
try:
await crud.create_run(db, run_data)
logger.info(f"Successfully created run record: {run_id}")
# Create benchmark results
created_results = []
logger.info(
f"Processing {len(upload_data.benchmark_results)} benchmark results"
)
for i, benchmark_result in enumerate(upload_data.benchmark_results, 1):
logger.debug(
f"Processing benchmark result {i}/{len(upload_data.benchmark_results)}: {benchmark_result.benchmark_name}"
)
# Convert worker format to internal format
stats_json = benchmark_result.stats_json
# Extract key metrics from the stats JSON
result_json = schemas.BenchmarkResultJson(
high_watermark_bytes=stats_json.get("metadata", {}).get(
"peak_memory", 0
),
allocation_histogram=[
[item["min_bytes"], item["count"]]
for item in stats_json.get("allocation_size_histogram", [])
],
total_allocated_bytes=stats_json.get("total_bytes_allocated", 0),
top_allocating_functions=[
schemas.TopAllocatingFunction(
function=alloc["location"],
count=alloc.get("count", 0),
total_size=alloc["size"],
)
for alloc in stats_json.get("top_allocations_by_size", [])[:10]
],
benchmark_name=benchmark_result.benchmark_name,
)
result_data = schemas.BenchmarkResultCreate(
run_id=run_id,
benchmark_name=benchmark_result.benchmark_name,
result_json=result_json,
flamegraph_html=benchmark_result.flamegraph_html,
)
try:
created_result = await crud.create_benchmark_result(db, result_data)
created_results.append(created_result.id)
logger.debug(
f"Successfully created benchmark result for {benchmark_result.benchmark_name}"
)
except Exception as e:
logger.error(
f"Failed to create benchmark result for {benchmark_result.benchmark_name}: {e}"
)
raise
logger.info(
f"Upload completed successfully: run_id={run_id}, created {len(created_results)} benchmark results"
)
# Clean up old flamegraphs if we have more than 100 runs for this environment
await cleanup_old_flamegraphs_if_needed(db, environment_id, max_flamegraphs=100)
# Clear any memray build failures for this binary+environment since upload succeeded
await db.execute(
delete(models.MemrayBuildFailure).where(
models.MemrayBuildFailure.binary_id == binary_id,
models.MemrayBuildFailure.environment_id == environment_id
)
)
await db.commit()
return {
"message": "Worker run uploaded successfully",
"run_id": run_id,
"commit_sha": commit_sha,
"binary_id": binary_id,
"environment_id": environment_id,
"results_created": len(created_results),
"result_ids": created_results,
}
except IntegrityError as e:
# Check for duplicate run constraint violation.
# asyncpg exposes constraint_name directly on the original exception;
# fall back to string matching for other backends (e.g. SQLite).
constraint = getattr(getattr(e, "orig", None), "constraint_name", None)
is_duplicate = (
constraint == "unique_commit_binary_env"
if constraint is not None
else "commit_sha" in str(e).lower()
and "binary_id" in str(e).lower()
and "environment_id" in str(e).lower()
)
if is_duplicate:
logger.error(
f"Upload failed: Duplicate run for commit {commit_sha[:8]}, binary '{binary_id}', environment '{environment_id}'"
)
raise HTTPException(
status_code=409,
detail=f"A run already exists for commit {commit_sha[:8]} with binary '{binary_id}' and environment '{environment_id}'. Duplicate uploads are not allowed.",
)
else:
logger.error(f"Database integrity error during upload: {e}")
raise HTTPException(
status_code=500, detail=f"Database integrity error: {str(e)}"
)
except HTTPException:
# Re-raise HTTP exceptions (validation errors) as-is
raise
except Exception as e:
logger.error(f"Unexpected error during upload processing: {e}")
raise HTTPException(
status_code=500, detail=f"Failed to upload worker run: {str(e)}"
)
@router.post("/report-memray-failure", response_model=dict)
async def report_memray_failure(
failure_data: schemas.MemrayFailureReport,
db: AsyncSession = Depends(get_database),
current_token: models.AuthToken = Depends(get_current_token),
):
"""Report a memray build failure for a specific commit and environment."""
logger = logging.getLogger(__name__)
logger.info(
f"Authenticated memray failure report from token '{current_token.name}' for commit {failure_data.commit_sha[:8]}, "
f"binary_id='{failure_data.binary_id}', environment_id='{failure_data.environment_id}'"
)
# Validate binary exists
binary = await crud.get_binary_by_id(db, binary_id=failure_data.binary_id)
if not binary:
logger.error(f"Memray failure report failed: Binary '{failure_data.binary_id}' not found")
raise HTTPException(
status_code=400,
detail=f"Binary '{failure_data.binary_id}' not found."
)
# Validate environment exists
environment = await crud.get_environment_by_id(db, environment_id=failure_data.environment_id)
if not environment:
logger.error(f"Memray failure report failed: Environment '{failure_data.environment_id}' not found")
raise HTTPException(
status_code=400,
detail=f"Environment '{failure_data.environment_id}' not found."
)
# Check if this failure is newer than any existing failure for this binary+environment
existing_failure = await db.execute(
select(models.MemrayBuildFailure)
.where(
models.MemrayBuildFailure.binary_id == failure_data.binary_id,
models.MemrayBuildFailure.environment_id == failure_data.environment_id
)
)
existing_failure = existing_failure.scalars().first()
if existing_failure:
# Check if the new failure is from a newer commit
if failure_data.commit_timestamp <= existing_failure.commit_timestamp:
logger.info(
f"Ignoring memray failure for {failure_data.commit_sha[:8]} as it's not newer than existing failure"
)
return {
"message": "Memray failure ignored (not newer than existing failure)",
"commit_sha": failure_data.commit_sha,
"binary_id": failure_data.binary_id,
"environment_id": failure_data.environment_id,
}
# Update existing failure with newer information
existing_failure.commit_sha = failure_data.commit_sha
existing_failure.error_message = failure_data.error_message
existing_failure.failure_timestamp = datetime.now()
existing_failure.commit_timestamp = failure_data.commit_timestamp
try:
await db.commit()
logger.info(f"Updated existing memray failure record for binary '{failure_data.binary_id}', environment '{failure_data.environment_id}'")
except Exception as e:
logger.error(f"Failed to update memray failure record: {e}")
raise HTTPException(
status_code=500, detail=f"Failed to update memray failure record: {str(e)}"
)
else:
# Create new failure record
failure_record = models.MemrayBuildFailure(
commit_sha=failure_data.commit_sha,
binary_id=failure_data.binary_id,
environment_id=failure_data.environment_id,
error_message=failure_data.error_message,
failure_timestamp=datetime.now(),
commit_timestamp=failure_data.commit_timestamp
)
try:
db.add(failure_record)
await db.commit()
logger.info(f"Created new memray failure record for commit {failure_data.commit_sha[:8]}")
except Exception as e:
logger.error(f"Failed to create memray failure record: {e}")
raise HTTPException(
status_code=500, detail=f"Failed to create memray failure record: {str(e)}"
)
return {
"message": "Memray failure reported successfully",
"commit_sha": failure_data.commit_sha,
"binary_id": failure_data.binary_id,
"environment_id": failure_data.environment_id,
}