Node monitoring
Enable the Nitro metrics server, scrape Prometheus metrics, monitor key metrics, and run health checks
Nitro exposes Prometheus-compatible metrics through a dedicated metrics server. Each Nitro binary reports its health through its own mechanism.
Metrics server
Enable the metrics server
Pass the --metrics flag when you start the node. Configure the server with these flags:
| Flag | Default | Description |
|---|---|---|
--metrics-server.addr | 127.0.0.1 | Listen address for the metrics server |
--metrics-server.port | 6070 | Listen port for the metrics server |
--metrics-server.update-interval | 3s | How often Nitro refreshes internal metrics |
Example:
nitro --metrics \
--metrics-server.addr 0.0.0.0 \
--metrics-server.port 6070Prometheus scrape endpoint
After you enable metrics, Nitro exposes a Prometheus-compatible endpoint at:
http://<metrics-server.addr>:<metrics-server.port>/debug/metrics/prometheuspprof runs as a separate server at /debug/pprof/ for CPU and memory profiling. Enable it with --pprof, and configure it with --pprof-cfg.addr and --pprof-cfg.port (default 127.0.0.1:6071). The metrics port does not serve pprof.
Key metrics to monitor
Group your alerts by concern. The metric names below are Nitro's Prometheus metric names.
| Concern | Metric | What it tells you |
|---|---|---|
| Memory | container_memory_rss | Actual container RSS (resident memory). Alert on this, not Go heap. Most Nitro memory is invisible to Go. |
| Sequencer feed | arb_feed_backlog_messages | Messages in the feed backlog. A growing backlog means the is not posting message batches to the . |
| Sequencing | arb_sequencer_backlog | Backlog at the . Sustained growth means the sequencer can't keep up with load. |
| Batch posting | arb_batchposter_wallet_eth | Batch poster wallet balance. If the wallet runs dry, batches stop posting. |
| Batch posting | arb_batchposter_suggestedtipcap | Suggested tip cap the batch poster pays. It tracks parent chain fee pressure. |
| RPC protection | arb_rpc_limitcheck_failure | Increments when the memory free-limit throttle rejects an RPC call (HTTP 429). |
Always alert on container_memory_rss, not Go heap metrics. Most of Nitro's memory is invisible to Go's runtime. To learn why, see Memory allocators in Nitro.
Health checks
Each Nitro component exposes health information through a mechanism specific to its binary.
Main Nitro node
The main nitro binary does not expose a dedicated /health endpoint. Infer node health from RPC availability. If the HTTP RPC port (default 8547) returns HTTP 200 for a valid JSON-RPC request, the node is healthy.
Example liveness check:
curl -sf -X POST http://localhost:8547 \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","method":"eth_chainId","params":[],"id":1}'Auxiliary components
- filterer: Exposes a
/livenessendpoint for health checks. - DA server: Exposes a
/healthendpoint.
Kubernetes
If you run Nitro on Kubernetes, the community Helm chart sets up the ServiceMonitor, startup and liveness probes, and a Grafana dashboard for you. To enable metrics scraping and configure probes on Kubernetes, see Enable monitoring and Probes in the Helm guide.
How is this guide?
Node tuning
Memory allocators, cache configuration, the CLI flags that control memory usage, GOMEMLIMIT, and validator tuning for Nitro nodes
Configuration system
How Nitro loads configuration from files, environment variables, CLI flags, and S3 — including precedence rules and how to dump running config