Starting, Stopping and Monitoring

Use xian-cli for operator-facing node lifecycle. Use xian-stack directly for backend debugging and xian-deploy for remote Linux hosts.

Lifecycle Commands

xian node start validator-1
xian node status validator-1
xian node endpoints validator-1
xian node health validator-1
xian node stop validator-1

Enabled services in the node profile start and stop with the node, including the dashboard, monitoring, BDS, IntentKit, DEX automation, and the shielded relayer.

status combines local artifacts, backend state, image information, and live RPC data. health is the concise machine-readable readiness view. endpoints prints the effective host URLs and ports, including Docker remapping.

For a broader preflight:

xian doctor validator-1
xian doctor validator-1 --skip-live-checks

The offline form validates artifacts without requiring a running node.

What to Monitor

At minimum, alert on:

• latest height and block age
• catching_up and peer count
• CometBFT and Xian process availability
• validator signing and missed-block behavior
• disk space and retained history
• application errors and transaction rejection rate
• BDS indexed height, lag, queue, pool, and spool when enabled
• parallel-execution acceptance and fallback ratios when enabled

An RPC endpoint that answers can still be unhealthy if height or block age is stale. Compare against trusted peers.

Logs

Xian application logs are separate from CometBFT logs. Rotated app logs live under the node home's xian/logs directory and follow the app_log_* settings in xian.toml.

tail -f ~/.cometbft/xian/logs/*.log

Start at INFO, use DEBUG for focused diagnosis, and use TRACE only briefly. Keep transaction_trace_logging disabled during normal operation.

Useful stages include check_tx, prepare_proposal, process_proposal, finalize_*, and simulate_tx.

Monitoring Surfaces

Surface	Use
CometBFT RPC	canonical node, block, transaction, and consensus reads
direct ABCI query	canonical Xian state and runtime queries
Xian and CometBFT metrics	alerts and time-series monitoring
dashboard REST/WebSocket	operator and explorer UI
BDS queries	indexed history and event feeds
GraphQL	optional generated query layer over BDS

Dashboard, GraphQL, and BDS are outside consensus. BDS can lag the finalized head while it catches up.

Local Backend Debugging

From xian-stack:

python3 ./scripts/backend.py start --no-bds-enabled --no-dashboard --no-monitoring
python3 ./scripts/backend.py status --no-bds-enabled --no-dashboard --no-monitoring
python3 ./scripts/backend.py endpoints --no-bds-enabled --no-dashboard --no-monitoring
python3 ./scripts/backend.py health --no-bds-enabled --no-dashboard --no-monitoring
python3 ./scripts/backend.py stop --no-bds-enabled --no-dashboard --no-monitoring

The backend is a machine-facing contract. Prefer xian node ... for routine operator work.

Public Monitoring

The stack binds query and monitoring services to loopback by default. Public exposure requires explicit enablement and appropriate TLS, authentication, firewalling, and rate limits. Do not publish Prometheus admin endpoints or raw node RPC simply to make a dashboard reachable.

Recovery and Maintenance

Use the page that owns the procedure:

• Recovery Plans for snapshot, state-sync, and emergency recovery choices
• Pruning and Retention for history and disk policy
• Protocol Governance and State Patches for governed forward patches
• BDS Indexed Queries for indexed consistency and query behavior
• 5-Validator Localnet E2E for whole-stack validation