Skip to main content

Documentation Index

Fetch the complete documentation index at: https://cyberpaisa-dof-mesh-40-27.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

MeshOrchestrator

core/mesh_orchestrator.py — singleton that manages task routing, node selection, and horizontal scaling.
from core.mesh_orchestrator import MeshOrchestrator

orchestrator = MeshOrchestrator.get_instance()
result = await orchestrator.dispatch(
    task="Analyze this code for security issues",
    task_type="security",
    priority="HIGH"
)

OrchestrationResult

@dataclass
class OrchestrationResult:
    task_id: str
    task_type: str
    selected_node: str
    cost_optimal_node: str
    dispatched_to: str
    circuit_state: str    # "CLOSED" | "OPEN" | "HALF_OPEN"
    success: bool
    error: str | None
    latency_ms: float
print(result.task_id)           # "task-7f3a..."
print(result.selected_node)     # "minimax"
print(result.cost_optimal_node) # "cerebras-llama"
print(result.dispatched_to)     # "minimax"
print(result.circuit_state)     # "CLOSED"
print(result.success)           # True
print(result.latency_ms)        # 142.3

Auto-Scaling Algorithm

ScalingDecision is computed every cycle:
D_net = 0.6 × (queue_depth / 50) + 0.4 × (avg_latency / sla_ms)

if D_net > 1.0:  # scale_up
if D_net < 0.2:  # scale_down
else:            # hold
D_netDecision
> 1.0scale_up — add node capacity
0.2–1.0hold — current capacity sufficient
< 0.2scale_down — reduce idle capacity

Task Priority Queue

Tasks are processed in this order:
CRITICAL > HIGH > NORMAL > LOW
result = await orchestrator.dispatch(
    task="Emergency security check",
    task_type="security",
    priority="CRITICAL"
)

Circuit Breaker States

Each provider has its own circuit breaker:
CLOSED    → Normal operation
OPEN      → Provider failed — requests skipped
HALF_OPEN → Testing if provider recovered
Check current states:
states = orchestrator.get_circuit_states()
# {"minimax": "CLOSED", "groq": "CLOSED", "cerebras": "OPEN"}

Mesh Metrics

metrics = orchestrator.get_metrics()
print(metrics["nodes_active"])          # 11
print(metrics["circuit_breakers_open"]) # 0
print(metrics["queue_depth"])           # 0
print(metrics["avg_latency_ms"])        # 142.3

Mesh Nodes

Node table and TTL backoff

Autonomous Operations

Task routing and auto-scaling guide