Performance Analytics

Offering clear metrics to monitor task performance and resource activity.

Dashboard Analytics is a pivotal feature within the Skyops ecosystem, providing real-time insights into the network's performance, resource utilization and task execution. By centralizing all relevant data into an intuitive interface, the dashboard empowers users and resource providers with actionable insights, enabling informed decisions and optimal use of resources.

⬣ Real-Time Resource Monitoring

The Analytics Dashboard offers comprehensive visibility into the performance and status of all active nodes in the network.

1/7

Code Example for Monitoring Resources:

# Check the status of all nodes in the network
skyops monitor-nodes --real-time

# Monitor a specific node's metrics
skyops node-status --node-id NODE1234

Example Scenario: A node owner uses the dashboard to monitor GPU usage and notices that their RTX 3060 is consistently underutilized. They adjust availability settings to match peak demand periods, increasing earnings.

⬣ Task Performance Insights

The dashboard provides detailed reports on task execution, allowing users to evaluate the efficiency and accuracy of AI workloads.

2/7

Code Example for Task Performance:

# Fetch task performance metrics
skyops task-metrics --task-id TASK5678

# Generate a performance report for completed tasks
skyops generate-report --time-range "2024-11-01 to 2024-11-19"

Example Use Case: A researcher submits a fine-tuning task for a GPT model. They track task duration, GPU utilization and validation accuracy in the dashboard, identifying areas for optimization in future tasks.

⬣ Earnings and Cost Analytics

The dashboard offers detailed financial insights for both resource providers and users.

For GPU Providers:

3/7

For Users:

4/7

Example Code for Financial Analytics:

# View total earnings as a provider
skyops earnings --provider-id PROVIDER789

# Analyze task costs as a user
skyops cost-analysis --user-id USER123

⬣ Network-Wide Metrics

The dashboard aggregates and visualizes data across the entire Skyops network, enabling macro-level insights for administrators and advanced users.

5/7

Example Visualization: The dashboard provides visual representations such as heatmaps showing node availability globally or line graphs tracking task completion rates over time.

Example Code for Network Metrics:

# Fetch network-wide statistics
skyops network-metrics --time-range "Last 24 Hours"

# Generate a heatmap of active nodes
skyops generate-heatmap --metric active_nodes

⬣ Dashboards and Visualization

The Analytics Dashboard offers user-friendly visualization tools to simplify data analysis. Some key visualization types include:

6/7

Visualization Example Code:

from skyops import Visualization

# Create a GPU usage graph
viz = Visualization()
viz.create_line_graph(metric="gpu_usage", time_range="last_week")

# Generate a heatmap of node activity
viz.create_heatmap(metric="active_nodes")

⬣ Benefits of the Analytics Dashboard

7/7

⬣ Code Example: Full Workflow

Scenario: A provider wants to monitor their node and track earnings, while a user wants to analyze task performance.

# Provider: Monitor node performance
skyops monitor-nodes --real-time

# Provider: Track earnings
sktops earnings --provider-id PROVIDER9876

# User: Analyze task performance
skyops task-metrics --task-id TASK4567

# User: Generate a report
skyops generate-report --time-range "2024-11-01 to 2024-11-11"

Performance Analytics is an essential tool for all participants in the Skyops ecosystem. By delivering real-time data, detailed insights and intuitive visualizations, it ensures efficient operation and optimal resource utilization for GPU providers and AI users alike.