Product Overview

A complete ops platform built for the demands of high-throughput Solana infrastructure.

Observability

Deep visibility into every layer of your Solana infrastructure. Track what matters when systems are under load.

  • - Real-time latency percentiles (P50, P95, P99) across all RPC endpoints
  • - Error rate tracking with automatic categorization and trend analysis
  • - Saturation metrics: connection pools, queue depths, resource utilization

How it works

1. Deploy lightweight agents alongside your RPC infrastructure
2. Agents emit structured telemetry to the Telemetry Plane
3. Real-time aggregation and anomaly detection surfaces issues instantly

Routing

Intelligent traffic management that adapts to real-time conditions and respects your operational policies.

  • - Policy-driven failover with customizable health checks and thresholds
  • - Endpoint scoring based on latency, availability, and historical performance
  • - Cost-aware routing to optimize spend without sacrificing reliability

How it works

1. Define routing policies in the Control Plane (latency SLOs, fallback rules, cost limits)
2. Data Plane scores endpoints in real-time based on telemetry and policy constraints
3. Traffic is automatically routed to optimal endpoints, with instant failover when needed

Claude Integrations

Optional copilot layer for incident analysis and config review. Runs only when requested.

runbook.generate

Draft incident runbooks from timeline + metrics window. Includes summary, suspected root cause, and next actions.

$ gate copilot runbook --incident 1842
Analyzing incident context...
Generated runbook draft saved to runbooks/inc-1842.md
config.review

Explain policy diffs and flag risky routing changes before apply. Reviews semantics, not just syntax.

$ gate copilot review policy/latency-failover.yaml
Analysis:
- Hysteresis window may be too short (60s)
- Consider increasing to 120s to prevent flapping
- Risk: medium
metrics.query

Natural-language queries over telemetry. Returns structured answers with metric context.

$ gate copilot query "why did P99 spike at 14:23?"
Analysis window: 14:15-14:35
Root cause: 3 EU endpoints saturated after config v1.8 deploy
Affected: eu-central-1, eu-west-1, eu-west-2
Recommendation: reduce weight on EU pool or revert v1.8

Copilot Credits

Claude integrations consume copilot credits (analysis units). Credits are billed per request and can be capped per workspace.

  • - Credits consumed only when explicitly requested via CLI or API
  • - All outputs logged with input context and model version
  • - No background agent or automatic inference
  • - Usage visible in billing dashboard with per-request breakdown
INCIDENT #1842
Severity: MediumStarted: 14:32 UTCResolved: 14:41 UTC
Summary:

Latency increase detected on 3 EU endpoints after config change.

Detected signals:
  • - P99 latency +47%
  • - Retry rate spike
  • - Endpoint saturation
Suggested actions:
  • - Reduce traffic weight on EU-2
  • - Roll back routing policy v1.8

Governance

Maintain control and compliance with comprehensive audit trails and safe change management.

  • - Immutable audit logs for all configuration changes and routing decisions
  • - Configuration versioning with rollback capabilities and diff visualization
  • - Safe rollout controls with canary deployments and automatic rollback on anomalies

Ready to get started?