petter2025 commited on
Commit
3854e5f
·
verified ·
1 Parent(s): a6671a5

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +14 -14
app.py CHANGED
@@ -134,18 +134,18 @@ def vacuum_db():
134
  logger.error(f"Vacuum failed: {e}")
135
 
136
  # ----------------------------------------------------------------------
137
- # Prometheus metrics (dummy if unavailable)
138
  # ----------------------------------------------------------------------
139
  if PROMETHEUS_AVAILABLE:
140
- decisions_total = Counter('arf_decisions_total', 'Total decisions made', ['action'])
141
- risk_gauge = Gauge('arf_current_risk', 'Current risk score')
142
- decision_latency = Histogram('arf_decision_latency_seconds', 'Time to evaluate intent')
143
- mcmc_runs = Counter('arf_mcmc_runs_total', 'Total MCMC runs')
144
  else:
145
- decisions_total = None
146
- risk_gauge = None
147
- decision_latency = None
148
- mcmc_runs = None
149
 
150
  # ----------------------------------------------------------------------
151
  # Thread‑safe history (in‑memory + DB backup)
@@ -168,8 +168,8 @@ def update_dashboard_data(decision: dict, risk: float):
168
  save_decision_to_db(decision, risk)
169
  # Update Prometheus metrics
170
  if PROMETHEUS_AVAILABLE:
171
- decisions_total.labels(action=decision.get("risk_level", "unknown")).inc()
172
- risk_gauge.set(risk)
173
 
174
  def refresh_history_from_db():
175
  """Load recent history from database (called at startup)."""
@@ -182,7 +182,7 @@ def refresh_history_from_db():
182
  decision_history.append((ts, dec, risk))
183
  risk_history.append((ts, risk))
184
  if PROMETHEUS_AVAILABLE:
185
- risk_gauge.set(risk) # update gauge with latest risk
186
 
187
  # ----------------------------------------------------------------------
188
  # Memory monitoring (daemon thread with graceful stop)
@@ -311,7 +311,7 @@ def handle_infra_with_governance(fault_type: str, context_window: int, session_s
311
  }
312
  # Record latency metric
313
  if PROMETHEUS_AVAILABLE:
314
- decision_latency.observe(time.time() - start_time)
315
  return output, session_state
316
  except Exception as e:
317
  logger.exception("Error in handle_infra_with_governance")
@@ -361,7 +361,7 @@ def run_hmc_mcmc(samples: int, warmup: int):
361
  samples = max(500, min(10000, int(samples)))
362
  warmup = max(100, min(2000, int(warmup)))
363
  if PROMETHEUS_AVAILABLE:
364
- mcmc_runs.inc() # record metric
365
 
366
  # Generate data: 10 observations with mean 0.5, std 0.2
367
  np.random.seed(42) # for reproducibility
 
134
  logger.error(f"Vacuum failed: {e}")
135
 
136
  # ----------------------------------------------------------------------
137
+ # Prometheus metrics (renamed to avoid conflict with Gradio components)
138
  # ----------------------------------------------------------------------
139
  if PROMETHEUS_AVAILABLE:
140
+ prom_decisions_total = Counter('arf_decisions_total', 'Total decisions made', ['action'])
141
+ prom_risk_gauge = Gauge('arf_current_risk', 'Current risk score')
142
+ prom_decision_latency = Histogram('arf_decision_latency_seconds', 'Time to evaluate intent')
143
+ prom_mcmc_runs = Counter('arf_mcmc_runs_total', 'Total MCMC runs')
144
  else:
145
+ prom_decisions_total = None
146
+ prom_risk_gauge = None
147
+ prom_decision_latency = None
148
+ prom_mcmc_runs = None
149
 
150
  # ----------------------------------------------------------------------
151
  # Thread‑safe history (in‑memory + DB backup)
 
168
  save_decision_to_db(decision, risk)
169
  # Update Prometheus metrics
170
  if PROMETHEUS_AVAILABLE:
171
+ prom_decisions_total.labels(action=decision.get("risk_level", "unknown")).inc()
172
+ prom_risk_gauge.set(risk)
173
 
174
  def refresh_history_from_db():
175
  """Load recent history from database (called at startup)."""
 
182
  decision_history.append((ts, dec, risk))
183
  risk_history.append((ts, risk))
184
  if PROMETHEUS_AVAILABLE:
185
+ prom_risk_gauge.set(risk) # update gauge with latest risk
186
 
187
  # ----------------------------------------------------------------------
188
  # Memory monitoring (daemon thread with graceful stop)
 
311
  }
312
  # Record latency metric
313
  if PROMETHEUS_AVAILABLE:
314
+ prom_decision_latency.observe(time.time() - start_time)
315
  return output, session_state
316
  except Exception as e:
317
  logger.exception("Error in handle_infra_with_governance")
 
361
  samples = max(500, min(10000, int(samples)))
362
  warmup = max(100, min(2000, int(warmup)))
363
  if PROMETHEUS_AVAILABLE:
364
+ prom_mcmc_runs.inc() # record metric
365
 
366
  # Generate data: 10 observations with mean 0.5, std 0.2
367
  np.random.seed(42) # for reproducibility