Documentation

Technical documentation for Gate platform.

Quickstart

Installation

Gate agents deploy alongside your existing RPC infrastructure. The deployment process takes less than 10 minutes and requires minimal configuration.

Step 1: Configure endpoints

Define your RPC endpoints in the Control Plane. Specify health check endpoints, expected latency baselines, and any custom metadata.

Step 2: Deploy agents

Install Gate agents as sidecar containers or standalone processes. Agents automatically begin emitting telemetry to the Telemetry Plane.

Step 3: Set routing policies

Create your first routing policy to define how traffic should flow. Start with a simple latency-based policy and refine based on observed behavior.

Core Concepts

The Four Planes

Gate architecture consists of four specialized planes: Telemetry (observability), Control (configuration), Data (routing), and AI (intelligence). Each plane has a focused responsibility and communicates through well-defined interfaces.

Endpoints

An endpoint represents a single RPC service instance. Endpoints are continuously scored based on health metrics, latency percentiles, and error rates. Scoring determines routing priorities.

Policies

Policies are declarative rules that govern routing behavior. They specify conditions (e.g., P99 latency > 200ms) and actions (e.g., shift traffic to fallback pool). Policies are versioned and auditable.

Incidents

When anomalies are detected, Gate creates an incident with full context: affected endpoints, metrics snapshots, and timeline. The AI Plane analyzes incidents and suggests remediation.

API Reference

Authentication

API requests require an API key passed in the Authorization header. Keys are scoped to specific permissions and can be time-limited. Rotate keys regularly.

Rate limits

API requests are rate-limited to 1000 requests per minute per API key. Burst allowance of 100 requests. Rate limit headers included in responses.

Example Request
GET /api/v1/endpoints
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Policy Examples

Latency-based routing
policy:
name: latency_failover
trigger:
metric: p99_latency
threshold: 200ms
action: shift_traffic