Spaces:
Sleeping
Sleeping
File size: 18,162 Bytes
4b9d59b | 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 | """
Data Consistency Insights for AI Debug System
Analyzes data consistency across distributed components and detects:
- Data flow across distributed nodes
- State divergence (same data, different states)
- Replication completion verification
- Data synchronization issues
Example insights:
- "Data sent to 5 nodes, 4 confirmed, 1 pending"
- "Data X saved on Node 1 but not Node 2"
- "Replication lag detected across 3 nodes"
"""
import asyncio
from collections import defaultdict
from datetime import datetime, timedelta
from typing import Any, Dict, List, Optional, Set, Tuple
from sqlalchemy.orm import Session
from sqlalchemy import and_, or_
from core.models import (
DebugEvent,
DebugInsight,
DebugStateSnapshot,
DebugInsightType,
DebugInsightSeverity,
)
from core.structured_logger import StructuredLogger
class ConsistencyInsightGenerator:
"""
Generates insights about data consistency across distributed components.
Tracks data flow through the system and detects inconsistencies,
replication delays, and synchronization issues.
"""
def __init__(self, db_session: Session):
"""
Initialize consistency insight generator.
Args:
db_session: SQLAlchemy database session
"""
self.logger = StructuredLogger(__name__)
self.db = db_session
async def analyze_data_flow(
self,
operation_id: str,
component_ids: List[str],
component_type: str = "agent",
) -> Optional[DebugInsight]:
"""
Analyze data flow across distributed components.
Tracks how data propagates through the system and identifies
any components that haven't received or acknowledged the data.
Args:
operation_id: Operation to analyze
component_ids: Expected components in the flow
component_type: Type of components
Returns:
Consistency insight or None
"""
try:
# Find all state snapshots for this operation
snapshots = (
self.db.query(DebugStateSnapshot)
.filter(
and_(
DebugStateSnapshot.operation_id == operation_id,
DebugStateSnapshot.component_type == component_type,
)
)
.order_by(DebugStateSnapshot.captured_at)
.all()
)
if not snapshots:
return None
# Track which components have snapshots
components_with_snapshots = set(s.component_id for s in snapshots)
# Find missing components
missing_components = set(component_ids) - components_with_snapshots
if missing_components:
# Data hasn't reached all components yet
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.WARNING.value,
title="Incomplete data propagation",
description=f"Data sent to {len(component_ids)} nodes but only "
f"{len(components_with_snapshots)} confirmed receipt",
summary=f"Data sent to {len(component_ids)} nodes, {len(missing_components)} pending",
evidence={
"operation_id": operation_id,
"expected_components": component_ids,
"confirmed_components": list(components_with_snapshots),
"missing_components": list(missing_components),
"propagation_rate": len(components_with_snapshots) / len(component_ids),
},
confidence_score=0.95,
suggestions=[
f"Check connectivity to missing nodes: {', '.join(missing_components)}",
"Verify message queues are processing",
"Check for network partitions",
],
scope="distributed",
affected_components=[
{"type": component_type, "id": comp_id} for comp_id in component_ids
],
generated_at=datetime.utcnow(),
)
# All components have data - check timing consistency
capture_times = [s.captured_at for s in snapshots if s.captured_at]
if len(capture_times) > 1:
min_time = min(capture_times)
max_time = max(capture_times)
replication_lag = (max_time - min_time).total_seconds()
if replication_lag > 5.0: # 5 second threshold
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.WARNING.value,
title="Replication lag detected",
description=f"Data replicated to all {len(component_ids)} nodes but with "
f"{replication_lag:.1f}s delay between first and last",
summary=f"Replication lag of {replication_lag:.1f}s across {len(component_ids)} nodes",
evidence={
"operation_id": operation_id,
"replication_lag_seconds": replication_lag,
"first_confirmation": min_time.isoformat() if min_time else None,
"last_confirmation": max_time.isoformat() if max_time else None,
},
confidence_score=0.90,
suggestions=[
"Investigate slow components",
"Check network latency between nodes",
"Review resource utilization",
],
scope="distributed",
affected_components=[
{"type": component_type, "id": comp_id} for comp_id in component_ids
],
generated_at=datetime.utcnow(),
)
# All good - data consistent
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.INFO.value,
title="Data consistent across all nodes",
description=f"Data successfully propagated to all {len(component_ids)} components",
summary=f"Data sent to {len(component_ids)} nodes, all confirmed",
evidence={
"operation_id": operation_id,
"component_count": len(component_ids),
"replication_complete": True,
},
confidence_score=1.0,
scope="distributed",
affected_components=[
{"type": component_type, "id": comp_id} for comp_id in component_ids
],
generated_at=datetime.utcnow(),
)
except Exception as e:
self.logger.error(
"Failed to analyze data flow",
operation_id=operation_id,
error=str(e),
)
return None
async def detect_state_divergence(
self,
operation_id: str,
component_type: str = "agent",
) -> Optional[DebugInsight]:
"""
Detect state divergence across components.
Identifies when the same data has different states on different nodes,
indicating a synchronization issue.
Args:
operation_id: Operation to analyze
component_type: Type of components
Returns:
Consistency insight or None
"""
try:
# Get all state snapshots for this operation
snapshots = (
self.db.query(DebugStateSnapshot)
.filter(
and_(
DebugStateSnapshot.operation_id == operation_id,
DebugStateSnapshot.component_type == component_type,
)
)
.order_by(DebugStateSnapshot.captured_at.desc())
.all()
)
if len(snapshots) < 2:
return None # Need at least 2 components to compare
# Get latest snapshot for each component
latest_snapshots = {}
for snapshot in snapshots:
if snapshot.component_id not in latest_snapshots:
latest_snapshots[snapshot.component_id] = snapshot
# Compare states across components
inconsistencies = self._compare_states(latest_snapshots)
if inconsistencies:
component_ids = list(latest_snapshots.keys())
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.CRITICAL.value,
title="State divergence detected",
description=f"Found {len(inconsistencies)} state inconsistencies across "
f"{len(component_ids)} components",
summary=f"Data differs across {len(component_ids)} components for {len(inconsistencies)} keys",
evidence={
"operation_id": operation_id,
"inconsistencies": inconsistencies,
"affected_keys": list(inconsistencies.keys()),
},
confidence_score=0.92,
suggestions=[
"Review synchronization logic",
"Check for concurrent modifications",
"Verify conflict resolution mechanisms",
"Investigate data partitioning issues",
],
scope="distributed",
affected_components=[
{"type": component_type, "id": comp_id} for comp_id in component_ids
],
generated_at=datetime.utcnow(),
)
return None
except Exception as e:
self.logger.error(
"Failed to detect state divergence",
operation_id=operation_id,
error=str(e),
)
return None
async def verify_replication_completion(
self,
operation_id: str,
expected_replicas: int,
component_type: str = "agent",
) -> Optional[DebugInsight]:
"""
Verify that data has been replicated to all expected nodes.
Args:
operation_id: Operation to verify
expected_replicas: Number of expected replicas
component_type: Type of components
Returns:
Consistency insight or None
"""
try:
# Count unique components that have data
component_count = (
self.db.query(DebugStateSnapshot)
.filter(
and_(
DebugStateSnapshot.operation_id == operation_id,
DebugStateSnapshot.component_type == component_type,
)
)
.distinct(DebugStateSnapshot.component_id)
.count()
)
if component_count < expected_replicas:
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.WARNING.value,
title="Incomplete replication",
description=f"Expected {expected_replicas} replicas but only {component_count} confirmed",
summary=f"{expected_replicas - component_count} replicas missing",
evidence={
"operation_id": operation_id,
"expected_replicas": expected_replicas,
"actual_replicas": component_count,
"completion_rate": component_count / expected_replicas,
},
confidence_score=0.98,
suggestions=[
"Check replication logs",
"Verify network connectivity",
"Review replication configuration",
],
scope="distributed",
affected_components=[{"type": component_type}],
generated_at=datetime.utcnow(),
)
# Replication complete
return DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.INFO.value,
title="Replication complete",
description=f"Data successfully replicated to all {expected_replicas} nodes",
summary=f"All {expected_replicas} replicas confirmed",
evidence={
"operation_id": operation_id,
"replica_count": component_count,
},
confidence_score=1.0,
scope="distributed",
affected_components=[{"type": component_type}],
generated_at=datetime.utcnow(),
)
except Exception as e:
self.logger.error(
"Failed to verify replication completion",
operation_id=operation_id,
error=str(e),
)
return None
async def analyze_sync_patterns(
self,
time_range: str = "last_1h",
) -> List[DebugInsight]:
"""
Analyze synchronization patterns across all operations.
Identifies systemic issues with data synchronization.
Args:
time_range: Time range to analyze
Returns:
List of consistency insights
"""
try:
insights = []
time_filter = self._parse_time_range(time_range)
# Find operations with incomplete replication
operations = (
self.db.query(DebugStateSnapshot.operation_id)
.filter(DebugStateSnapshot.captured_at >= time_filter)
.group_by(DebugStateSnapshot.operation_id)
.all()
)
for (operation_id,) in operations:
# Count components per operation
component_count = (
self.db.query(DebugStateSnapshot)
.filter(
and_(
DebugStateSnapshot.operation_id == operation_id,
DebugStateSnapshot.captured_at >= time_filter,
)
)
.distinct(DebugStateSnapshot.component_id)
.count()
)
# If an operation has only 1 component, it might not be replicating
if component_count == 1:
insights.append(
DebugInsight(
insight_type=DebugInsightType.CONSISTENCY.value,
severity=DebugInsightSeverity.INFO.value,
title="Single-component operation detected",
description=f"Operation {operation_id} has data on only 1 component",
summary=f"Operation may not be configured for replication",
evidence={"operation_id": operation_id},
confidence_score=0.70,
suggestions=[
"Verify replication is configured",
"Check if this is a local-only operation",
],
scope="distributed",
affected_components=[],
generated_at=datetime.utcnow(),
)
)
return insights
except Exception as e:
self.logger.error("Failed to analyze sync patterns", error=str(e))
return []
def _compare_states(
self,
snapshots: Dict[str, DebugStateSnapshot],
) -> Dict[str, Any]:
"""
Compare states across multiple components.
Args:
snapshots: Component ID -> Snapshot mapping
Returns:
Dictionary of inconsistent keys and their values
"""
inconsistencies = {}
# Get all keys from the first snapshot
first_snapshot = list(snapshots.values())[0]
all_keys = set(first_snapshot.state_data.keys()) if first_snapshot.state_data else set()
# For each key, compare values across all snapshots
for key in all_keys:
values = {}
for comp_id, snapshot in snapshots.items():
if snapshot.state_data and key in snapshot.state_data:
values[comp_id] = snapshot.state_data[key]
# Check if all values are the same
unique_values = set(str(v) for v in values.values())
if len(unique_values) > 1:
inconsistencies[key] = {
"values": values,
"divergence_detected": True,
}
return inconsistencies
def _parse_time_range(self, time_range: str) -> datetime:
"""Parse time range string to datetime."""
now = datetime.utcnow()
if time_range == "last_1h":
return now - timedelta(hours=1)
elif time_range == "last_24h":
return now - timedelta(hours=24)
elif time_range == "last_7d":
return now - timedelta(days=7)
else:
return now - timedelta(hours=1)
|