How to use D-Miner for cloud-based monitoring? (Dashboard)

Dashboard Overview

1. The D-Miner dashboard provides a real-time visual interface for tracking mining performance across distributed cloud nodes. It aggregates hash rate, temperature, power consumption, and uptime metrics into unified panels.

2. Each cloud miner instance appears as a distinct tile with color-coded status indicators—green for active, yellow for throttled, red for offline or error states.

3. Time-series charts render historical data at 1-minute, 15-minute, and 1-hour granularities, enabling rapid identification of performance dips or thermal anomalies.

4. Customizable widgets allow users to pin specific metrics such as rejected share count, pool latency, or GPU utilization per node.

5. Role-based access controls restrict dashboard visibility—administrators see full infrastructure topology while operators view only assigned node groups.

Configuration Workflow

1. Cloud miners register automatically upon boot if preconfigured with valid API keys and endpoint URIs pointing to the central D-Miner service.

2. Manual registration requires entering the cloud instance’s public IP, SSH credentials, and mining software version into the “Add Node” form.

3. Dashboard auto-detects hardware specs—including GPU model, VRAM capacity, and PCIe bandwidth—and maps them to optimized mining profiles.

4. Users assign tags like “eth-mainnet”, “ergo-testnet”, or “low-priority” to filter nodes during bulk operations or alert routing.

5. Configuration changes—such as switching pools or adjusting intensity settings—are pushed via signed JSON payloads and logged in immutable audit trails.

Data Synchronization Protocol

1. D-Miner uses a lightweight MQTT-based telemetry layer that transmits encrypted metric packets every 3 seconds over TLS 1.3.

2. Each packet includes a monotonic sequence number, timestamp, and SHA-256 checksum to prevent replay or tampering attacks.

3. Cloud instances with intermittent connectivity buffer up to 10 minutes of metrics locally and transmit them in-order upon reconnection.

4. The dashboard displays synchronization latency as a delta between device clock and NTP-synchronized server time—values exceeding 500ms trigger a warning icon.

5. All raw telemetry data is stored in immutable Parquet files on object storage, accessible only via time-bound presigned URLs.

Alerting and Notification Engine

1. Threshold rules are defined per metric—for example, “GPU temp > 85°C for > 90 seconds” or “hash rate drop > 40% over 5 minutes”.

2. Alert conditions support boolean logic: (temp > 85 AND fan_speed 0.02).

3. Notifications dispatch simultaneously to Slack webhooks, PagerDuty incidents, and email gateways with embedded dashboard permalink snapshots.

4. Acknowledged alerts remain visible in the dashboard’s “Resolved” tab for 7 days with full context including related logs and configuration snapshots.

5. Critical alerts automatically pause mining processes on affected nodes until manual override or scheduled resume window.

Troubleshooting Common Dashboard Issues

Q1: Why does a newly added cloud node show “Unknown Status” after 5 minutes?It indicates failed handshake with the D-Miner backend—verify firewall rules permit outbound TCP 8883 to the configured MQTT broker and confirm the node’s system clock is within ±2 seconds of UTC.

Q2: Can I export raw sensor data from the dashboard for external analysis?Yes—click the “Export” button on any time-series chart to download CSV containing timestamps, metric names, values, and node IDs; exports are limited to 1 million rows per request.

Q3: Why do hash rate graphs display flatlined values for certain intervals?This occurs when the mining process was terminated or crashed; check the “Process Logs” tab for exit codes and restart timestamps—exit code 137 signals OOM kill by the cloud provider’s kernel.

Q4: Is it possible to disable automatic firmware updates triggered by dashboard actions?Yes—navigate to “Settings > Auto-Update Policy” and toggle off “Apply critical GPU firmware patches”; this setting persists across all enrolled nodes in the selected tag group.