Aurigraph's Node Health Scoring system continuously evaluates every node in the network using a 4-factor health score (0-100) based on latency, error rate, disk utilization, and memory pressure. Nodes are graded A through F, and the system automatically generates rebalance recommendations when nodes fall below configurable thresholds — ensuring network reliability without manual intervention.
Manual Monitoring:
Operator checks dashboards periodically | (Reactive — problems found after impact) Alerts based on single-metric thresholds | (No holistic view of node condition) Rebalancing decisions made ad-hoc | (Inconsistent, slow response)
Aurigraph Health Scoring:
4 metrics sampled continuously per node: |- Latency (consensus round-trip) |- Error rate (failed operations / total) |- Disk utilization (% used) |- Memory pressure (% heap used) | Weighted composite score (0-100): |- A (90-100): Excellent |- B (70-89): Good |- C (50-69): Acceptable |- D (30-49): Degraded |- F (0-29): Critical | Auto-rebalance recommendations: |- Generated when node drops below threshold |- Actionable: migrate load, restart, or decommission
+----------------------------------------------------------+ | Node Health Score Engine | +----------------------------------------------------------+ | | | Per-Node Metrics (sampled every 5 seconds): | | +---------------------------------------------------+ | | | Metric | Weight | Scoring Function | | | |---------------+--------+---------------------------| | | | Latency | 30% | 100 - (avgMs / threshold) | | | | Error Rate | 30% | 100 - (rate * 1000) | | | | Disk Usage | 20% | 100 - (pctUsed * 1.1) | | | | Memory | 20% | 100 - (heapPct * 1.1) | | | +---------------------------------------------------+ | | | | Composite Score: | | score = (latencyScore * 0.30) | | + (errorScore * 0.30) | | + (diskScore * 0.20) | | + (memoryScore * 0.20) | | | | Clamped to [0, 100], rounded to integer | +----------------------------------------------------------+
Node: validator-eu-1 Latency: 12ms avg (illustrative, not measured on production; design threshold 100ms) -> latencyScore = 88 Errors: 0.1% rate -> errorScore = 90 Disk: 45% used -> diskScore = 50.5 Memory: 38% heap -> memoryScore = 58.2 Composite = (88 * 0.30) + (90 * 0.30) + (50.5 * 0.20) + (58.2 * 0.20) = 26.4 + 27.0 + 10.1 + 11.64 = 75.14 -> Grade: B Node: business-ap-3 Latency: 250ms avg (illustrative, not measured on production; design threshold 100ms) -> latencyScore = 0 (clamped) Errors: 5.2% rate -> errorScore = 48 Disk: 92% used -> diskScore = 0 (clamped) Memory: 87% heap -> memoryScore = 4.3 Composite = (0 * 0.30) + (48 * 0.30) + (0 * 0.20) + (4.3 * 0.20) = 0 + 14.4 + 0 + 0.86 = 15.26 -> Grade: F
Grade Distribution: A |============== | 14 nodes (38%) B |=========== | 11 nodes (30%) C |====== | 6 nodes (16%) D |==== | 4 nodes (11%) F |== | 2 nodes (5%) +---------------+ 0 5 10 15 Network Health: 76.3 (Grade B) = weighted average of all node scores = sum(nodeScore * nodeWeight) / sum(nodeWeight) where nodeWeight = 1.0 for validators, 0.5 for business, 0.3 for EI
When a node's score drops below the configurable threshold (default: 50, Grade C), the system generates actionable recommendations:
Rebalance Decision Tree: Node score drops below threshold (default: 50) | +-- Latency dominant? (latencyScore < 30) | +-- Yes -> MIGRATE_LOAD | | reason: "High latency (XXms avg) — redistribute to lower-latency peers" | | target: [list of A-grade nodes in same region] | +-- Error rate dominant? (errorScore < 30) | +-- Yes -> MIGRATE_LOAD + investigation flag | | reason: "Error rate X.X% exceeds threshold — check logs" | +-- Memory dominant? (memoryScore < 20) | +-- Yes -> RESTART_NODE | | reason: "Heap at XX% — graceful restart recommended" | +-- Disk dominant? (diskScore < 20) | +-- Yes -> EXPAND_DISK | | reason: "Disk at XX% — provision additional storage" | +-- Score F for >10 minutes? +-- Yes -> DECOMMISSION | reason: "Node critically unhealthy for >10m — replace"
Health score changes and rebalance actions emit DMRV attestations:
NODE_HEALTH_SCORED
NODE_GRADE_CHANGED
REBALANCE_RECOMMENDED
REBALANCE_EXECUTED
NODE_DECOMMISSIONED
Node Health Scores (/api/v11/nodes/health-scores):
/api/v11/nodes/health-scores
# Get health scores for all nodes curl https://dlt.aurigraph.io/api/v11/nodes/health-scores \ -H "Authorization: Bearer {token}" # Response: # { # "networkHealth": 76.3, # "networkGrade": "B", # "nodes": [ # { # "nodeId": "validator-eu-1", # "score": 75, # "grade": "B", # "metrics": { # "latencyMs": 12, # "errorRate": 0.001, # "diskPct": 45, # "memoryPct": 38 # }, # "recommendation": null # }, # { # "nodeId": "business-ap-3", # "score": 15, # "grade": "F", # "metrics": { ... }, # "recommendation": { # "type": "DECOMMISSION", # "reason": "Node critically unhealthy for >10m", # "urgency": "HIGH" # } # } # ] # } # Get health score for a specific node curl https://dlt.aurigraph.io/api/v11/nodes/health-scores/{nodeId} # Get rebalance recommendations (nodes below threshold) curl https://dlt.aurigraph.io/api/v11/nodes/health-scores/recommendations?threshold=50 # Response: # { # "threshold": 50, # "degradedNodes": 6, # "recommendations": [ # { # "nodeId": "business-ap-3", # "score": 15, # "grade": "F", # "type": "DECOMMISSION", # "reason": "Node critically unhealthy for >10m — replace", # "targetNodes": ["business-ap-4", "business-ap-5"], # "urgency": "HIGH" # } # ] # }
healthScore.sampleIntervalMs
healthScore.latencyThresholdMs
healthScore.rebalanceThreshold
healthScore.decommissionTimeoutMin
healthScore.weights.latency
healthScore.weights.errorRate
healthScore.weights.disk
healthScore.weights.memory
Explore: About · Products · Features · Whitepapers · Sustainability · Tokenized carbon credits · Real-world asset tokenization · SDK · Blog · Plain-text site index · Sitemap · Contact: contact@aurigraph.io