Monitoring & Analytics

OmniSocial offers a Prometheus + Grafana-based monitoring stack to ensure reliability across its infrastructure, including relayers, the subgraph, and optional full nodes.

This guide outlines the setup and key metrics you should track to maintain performance and availability.

🧩 Monitoring Stack Overview

Tool	Purpose
Prometheus	Metrics collection and scraping
Grafana	Dashboard visualization
Health API	Local HTTP health checks

All services expose Prometheus-compatible endpoints.

🧪 Monitored Components

1. Relayer Nodes

If you operate a relayer (for meta-transactions or fee abstraction), monitor:

• ✅ Transaction submission success rate

• 📦 Queue backlog (pending tx count)

• 🧾 Signed tx per second

• ❗ Failures (e.g., insufficient funds, nonce too low)

Relayer Metrics Exported:

relayer_tx_count_total{status="success|fail"}
relayer_queue_length
relayer_gas_usage_avg

2. Subgraph Indexer

The subgraph exposes important sync and indexing metrics:

• 📊 Block synchronization lag

• 🔄 Reorg detection

• ⚠️ Failed handlers or dropped events

Key Metrics:

subgraph_block_lag
subgraph_latest_block
subgraph_processed_events_total

🛠 Note: You can also query the subgraph status via The Graph's /status endpoint:

{
  indexingStatusForCurrentVersion(subgraphName: "omnisocial/zk-mainnet") {
    synced
    health
    fatalError { message }
    chains {
      latestBlock { number }
      chainHeadBlock { number }
    }
  }
}

3. OmniSocial Web App

You may optionally monitor the dApp frontend for SSR errors, API rate limits, and network status. For production builds:

• Use Vercel Analytics or OpenTelemetry.

• For self-hosted versions, Prometheus can monitor:

frontend_http_requests_total
frontend_page_load_time_avg
frontend_graphql_query_latency

4. Wallet Relayer Services

If you're running the OmniWallet relayer, you can enable Prometheus metrics via a --metrics flag or .env setting:

ENABLE_METRICS=true
METRICS_PORT=9090

Exposes metrics like:

• Relayed txs per address

• Signature verification success/failure

• Gas cost per transaction

🧾 Health Check Endpoints

Each service exposes a basic health route:

Service	Endpoint	Response
Subgraph	/health	{"healthy":true}
Relayer	/status or /health	HTTP 200/500
Frontend	/api/health	pong

You can plug these into uptime monitoring tools like UptimeRobot, BetterStack, or Grafana OnCall.

📊 Grafana Dashboards

OmniSocial maintains default dashboards with the following panels:

• ⚙️ Infrastructure overview

• 🔁 Subgraph sync and block lag

• 📤 Relayer performance

• ⏱ GraphQL query latency

• 📈 Live system usage (DAOs, tips, posts per minute)

Dashboards can be imported from infra/monitoring/dashboards/*.json if running locally.

📌 Optional Setup Script

For self-hosted deployments, we provide:

infra/monitoring/docker-compose.yml

Which includes:

• Prometheus

• Grafana (port :3000, default creds admin/admin)

• Node exporter

• Alertmanager (optional)