Spaces:
Sleeping
Sleeping
File size: 12,330 Bytes
3ef7678 | 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 | """
Symptom Engine for Repair Strategy System V2.
HARD RULES ENFORCED:
1. Log ambiguity: No single log message uniquely identifies one root cause.
Every log maps to >= 2 root causes.
2. Metrics are derived from the cascade-affected symptomatic state.
"""
import random
from typing import Dict, List, Set
from models_v2 import LogEntry, LogSeverity, ComponentMetrics
# Hard Rule #1: Ambiguous overlapping log pool
# Every message MUST appear in the list for >= 2 distinct components.
# The keys represent the 'symptomatic' state of a component (which may be caused by a different root cause)
LOG_POOL: Dict[str, Dict[str, List[str]]] = {
"database": {
"down": [
"Connection timeout", # Shared with Cache down
"Service unavailable", # Shared with API down
"Upstream dependency failure", # Shared with Cache down
"Request queue backing up", # Shared with Queue overloaded
"High response time", # Shared with Queue overloaded
"Elevated read pressure", # Shared with Cache down
"Consumer lag detected", # Shared with Queue overloaded
"504 Gateway Timeout", # Shared with API degraded
"Slow slow-query log rising", # Shared with API degraded
"Cache miss rate critical", # Shared with Cache down
"Fatal IO error" # Shared with Queue down
]
},
"api": {
"down": [
"Service unavailable", # Shared with DB down
"Endpoint unhandled exception", # Shared with Cache down
"Process exited unexpectedly" # Shared with Queue overloaded
],
"degraded": [
"504 Gateway Timeout", # Shared with DB down
"Slow slow-query log rising", # Shared with DB down
"Thread pool exhaustion" # Shared with Queue overloaded
]
},
"cache": {
"down": [
"Connection timeout", # Shared with DB down
"Upstream dependency failure", # Shared with DB down
"Cache miss rate critical", # Shared with DB down (read pressure)
"Elevated read pressure", # Shared with DB down
"Memory pool exhausted", # Shared with Queue overloaded
"Queue backlog exceeding threshold", # Shared with Queue overloaded
"Endpoint unhandled exception", # Shared with API down
"Broker connection reset" # Shared with Queue down
]
},
"queue": {
"down": [
"Process exited unexpectedly", # Shared with API down
"Broker connection reset", # Shared with Cache down
"Fatal IO error" # Shared with DB down
],
"degraded": [ # Corresponds to "overloaded"
"High response time", # Shared with DB down
"Request queue backing up", # Shared with DB down
"Memory pool exhausted", # Shared with Cache down
"Queue backlog exceeding threshold", # Shared with Cache down
"Consumer lag detected", # Shared with DB down
"Thread pool exhaustion" # Shared with API degraded
]
}
}
# Healthy baseline metrics
BASE_METRICS = {
"database": {"latency": 30, "error_rate": 0.0, "queue_load": 0.0, "cpu_usage": 15.0},
"api": {"latency": 50, "error_rate": 0.0, "queue_load": 0.1, "cpu_usage": 20.0},
"cache": {"latency": 10, "error_rate": 0.0, "queue_load": 0.0, "cpu_usage": 30.0},
"queue": {"latency": 20, "error_rate": 0.0, "queue_load": 0.2, "cpu_usage": 10.0},
}
class SymptomEngine:
def __init__(self, seed: int):
self._rng = random.Random(seed)
def _get_log_variations(self, base_log: str) -> str:
variations = {
"Connection timeout": [
"Connection timeout on socket {}...",
"Err: socket hangup {}ms",
"Timeout Error: Target unreachable"
],
"Service unavailable": [
"Service unavailable (503) at path /{}",
"CRITICAL: Application unresponsive",
"HTTP 503 Service Temporarily Unavailable"
],
"Upstream dependency failure": [
"[DependencyError] Upstream failed to respond",
"Upstream dependency failure: {} retry",
"Failed to ping upstream node {}"
],
"Request queue backing up": [
"Request queue backing up, count: {}",
"WARNING: Queue depth exceeding safe limits",
"High request buffering ({})"
],
"High response time": [
"High response time > {}ms",
"Latency spike detected in APM: {}",
"Slow response: {}ms avg"
],
"Elevated read pressure": [
"Elevated read pressure on node {}",
"High read IOPS detected ({})",
"Warning: Read volume approaching limit"
],
"Consumer lag detected": [
"Consumer lag detected: {} units",
"Processing delay on consumer group {}",
"Lag spike: {} offset difference"
],
"504 Gateway Timeout": [
"504 Gateway Timeout on {}",
"Gateway timeout contacting upstream",
"Nginx 504: Target took too long"
],
"Slow slow-query log rising": [
"Slow slow-query log rising",
"Slow query detected: {} sec",
"DB Query execution time > {}ms"
],
"Cache miss rate critical": [
"Cache miss rate critical ({}%)",
"Warning: Cache hit/miss ratio inverted",
"Excessive cache misses on key prefix {}"
],
"Fatal IO error": [
"Fatal IO error at block {}",
"Disk IO exception volume mounted at {}",
"IOError: operation failed"
],
"Memory pool exhausted": [
"Memory pool exhausted (alloc: {}MB)",
"OOM Warning: pool size limit reached",
"Cannot allocate memory for {}"
],
"Queue backlog exceeding threshold": [
"Queue backlog exceeding threshold: {} reqs",
"Backlog critical: {} items",
"Spooling exceeded maximum partition limit"
],
"Endpoint unhandled exception": [
"Endpoint unhandled exception in route {}",
"Unhandled runtime Error at GET /{}",
"500 Internal Server Error: {}"
],
"Broker connection reset": [
"Broker connection reset by peer",
"TCP RST received from broker {}",
"Connection dropped: Broker unavailable"
]
}
pool = variations.get(base_log, [base_log])
choice = self._rng.choice(pool)
# Random template filling for realism
if "{}" in choice:
filler = self._rng.randint(10, 9999)
choice = choice.format(filler)
return choice
self._verify_ambiguity_constraint()
def _verify_ambiguity_constraint(self):
"""
Hard Rule verification at startup:
Ensures NO log exists that maps to exactly 1 root cause.
(We map to symptomatic components here, but the principle is that
different root causes trigger these symptoms, making the log ambiguous).
"""
message_to_sources: Dict[str, Set[str]] = {}
for comp, statuses in LOG_POOL.items():
for status, msgs in statuses.items():
for msg in msgs:
if msg not in message_to_sources:
message_to_sources[msg] = set()
message_to_sources[msg].add(comp)
violations = []
for msg, sources in message_to_sources.items():
if len(sources) < 2:
# Some logs map to different *statuses* of the same component,
# but to be truly ambiguous across root causes, we need strict overlap.
# The manual pool defines them carefully, but this safety check ensures
# future additions don't break the rule.
violations.append(f"'{msg}' only appears in {sources}")
# If the pool is perfectly designed, violations will be empty.
if violations:
raise ValueError(f"Ambiguity Constraint Violation! Logs uniquely identifying a component: {violations}")
def generate_metrics(self, state: Dict[str, Dict[str, any]]) -> Dict[str, ComponentMetrics]:
"""
Generate metrics based on the current (cascade-applied) state.
"""
metrics = {}
# 1. Base translation from status to metrics
for comp, data in state.items():
status = data["status"]
base = BASE_METRICS[comp]
m = {"latency": base["latency"], "error_rate": base["error_rate"],
"queue_load": base["queue_load"], "cpu_usage": base["cpu_usage"]}
if status == "down":
m["latency"] = 9999
m["error_rate"] = 1.0
m["cpu_usage"] = 0.0 # Process is dead
elif status == "degraded":
m["latency"] *= int(3 + self._rng.random() * 2)
m["error_rate"] = 0.3 + self._rng.random() * 0.2
m["cpu_usage"] = min(100.0, m["cpu_usage"] * 4)
metrics[comp] = m
# 2. Cross-component metric leaking (Symptoms)
# If DB is down, API naturally struggles
if state["database"]["status"] == "down" and state["api"]["status"] != "down":
metrics["api"]["latency"] += int(300 + self._rng.random() * 100)
metrics["api"]["error_rate"] = min(1.0, metrics["api"]["error_rate"] + 0.5)
metrics["api"]["cpu_usage"] = min(100.0, metrics["api"]["cpu_usage"] * 2)
# If Cache is down, Queue load naturally spikes due to DB pressure
if state["cache"]["status"] == "down":
metrics["queue"]["queue_load"] += 0.3 + self._rng.random() * 0.2
# Format output
formatted_metrics = {}
for comp, m in metrics.items():
formatted_metrics[comp] = ComponentMetrics(
latency=int(m["latency"]),
error_rate=round(m["error_rate"], 2),
queue_load=round(m["queue_load"], 2),
cpu_usage=round(m["cpu_usage"], 1)
)
return formatted_metrics
def generate_logs(self, state: Dict[str, Dict[str, any]], step: int) -> List[LogEntry]:
"""
Generate ambiguous logs based on active symptoms.
Returns 0-3 logs per broken component.
"""
logs = []
for comp, data in state.items():
status = data["status"]
if status == "healthy":
continue
pool = LOG_POOL.get(comp, {}).get(status, [])
if not pool:
# Fallback if specific status pool isn't defined
pool = LOG_POOL.get(comp, {}).get("down", ["System error"])
# Randomly select 1-2 logs for this symptom
num_logs = self._rng.randint(1, 2)
selected_msgs = self._rng.sample(pool, min(num_logs, len(pool)))
for msg in selected_msgs:
severity = LogSeverity.ERROR if status == "down" else LogSeverity.WARN
logs.append(LogEntry(
timestamp=step * 1000 + self._rng.randint(10, 990),
message=msg,
severity=severity
))
# Shuffle to prevent component ordering from being a tell
self._rng.shuffle(logs)
return logs
|