Operations

Running, monitoring, GPU configuration, and troubleshooting your ADI Rollup deployment.

Running Everything Together

To bring up the full stack in one command:

# Core + Infrastructure
docker compose \
  -f docker-compose.yml \
  -f docker-compose.infra.yml \
  up -d

# Core + Infrastructure + Blockscout
docker compose \
  -f docker-compose.yml \
  -f docker-compose.infra.yml \
  -f docker-compose.blockscout.yml \
  up -d

Rate limits and allowlisting

The mainnet RPC at rpc.adifoundation.ai enforces per-IP rate limits for L3 traffic.

If sequencer startup hits HTTP 429 with Cloudflare 1015, the issue is infrastructure configuration, not chain logic. Contact ADI Ops to allowlist your outbound IP.

Testnet at rpc.ab.testnet.adifoundation.ai does not enforce these limits.

If you are hitting 429 during startup, set restart: "no" on the affected services until your IP is allowlisted. Keep restart: unless-stopped for transient upstream failures such as 502.

---

Port Reference

Port	Service	Description
3050	Sequencer	L2 JSON-RPC endpoint
3051	External Node	L2 JSON-RPC (read replica)
3053	Sequencer	Block replay server used by the External Node
3054	External Node	Replay server override to avoid conflict on the same host
3320	Sequencer	Prover API (internal)
3000	Bridge	dApp Portal / Bridge UI
3010	Explorer App	Block Explorer frontend
3020	Explorer API	Block Explorer REST API
3040	Data Fetcher	Explorer data fetcher
4000	Blockscout	Blockscout frontend (optional)

---

GPU Configuration

Single GPU

If you have a dedicated GPU per prover, use the full GPU UUID:

nvidia-smi -L
# GPU 0: NVIDIA A100 80GB (UUID: GPU-12345678-abcd-...)

Set in .env:

GPU_DEVICE_FRI=GPU-12345678-abcd-...
GPU_DEVICE_SNARK=GPU-87654321-dcba-...

Scaling Provers

For higher proving throughput, add more prover instances. Each needs a unique name and Prometheus port:

  fri-prover-2:
    image: ${PROVER_FRI_IMAGE}
    container_name: ${CHAIN_SHORT_NAME}-fri-prover-2
    restart: unless-stopped
    network_mode: host
    depends_on: [server]
    environment:
      RUST_LOG: "info"
    volumes:
      - ./volumes/prover:/prover
    command:
      - "--base-url"
      - "http://127.0.0.1:3320"
      - "--app-bin-path"
      - "/multiblock_batch.bin"
      - "--enabled-logging"
      - "--prover-name"
      - "fri-prover-2"
      - "--prometheus-port"
      - "3125"
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ["${GPU_DEVICE_FRI_2}"]
              capabilities: [gpu]

---

Viewing Logs

# All services
docker compose logs -f

# Specific service
docker compose logs -f server
docker compose logs -f fri-prover
docker compose logs -f snark-prover

# Infrastructure stack
docker compose -f docker-compose.infra.yml logs -f explorer-worker

---

Monitoring

Endpoint	What It Shows
http://localhost:3050	Sequencer JSON-RPC
http://localhost:3051	External Node JSON-RPC
http://localhost:3312/metrics	Sequencer Prometheus metrics
http://localhost:3316/metrics	External Node Prometheus metrics
http://localhost:3124/metrics	FRI Prover Prometheus metrics
http://localhost:3126/metrics	SNARK Prover Prometheus metrics

Check sync status:

# Sequencer block number
curl -s -X POST http://localhost:3050 \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"eth_blockNumber","params":[],"id":1}'

# GPU utilization
watch -n 1 nvidia-smi

---

Restarting Services

# Restart a single service
docker compose restart server

# Restart the entire core stack
docker compose down && docker compose up -d

---

Troubleshooting

Error	Likely cause	Resolution
failed to fetch finalized L1 state with 429	Your settlement RPC is rate-limiting L3 traffic	Check the upstream response. On mainnet, contact ADI Ops to allowlist your outbound IP. Set restart: "no" until allowlisting is complete.
failed to fetch finalized L1 state with 502	Upstream settlement RPC is unavailable or returning a transient gateway error	Keep restart: unless-stopped. Retry after the upstream recovers, or switch to a healthier RPC endpoint.
sender does not have enough funds	An operator wallet is missing settlement-layer gas funds	Fund the commit, prove, and execute operator wallets with the settlement-layer gas token. If base_token_address was omitted at deploy time, that token is ADI.
Existing batch first block does not match	Local node state does not match the chain you are trying to run	Stop the node. Remove the affected node's RocksDB and replay data. Keep reusable wallet files only. Do not reuse genesis.yaml or contracts.yaml across chains.
L1 commit watcher component failed	Settlement RPC is wrong, unavailable, rate-limited, or paired with the wrong settlement chain ID	Verify L1_RPC_URL reaches your settlement layer. Confirm the generated genesis.json has the correct l1_chain_id. For ADI, use 36900 on mainnet or 99999 on testnet. If you use a public RPC, relax the watcher settings to l1_watcher_poll_interval: "2s" and l1_watcher_max_blocks_to_process: 10.