IoT-Edge Telemetry Dataset for Intelligent Task Offloading and AI Workload Placement

Description This dataset contains time-series telemetry data collected from a Kubernetes-based IoT–Edge computing environment instrumented with Prometheus and Node Exporter. The setup consists of two IoT nodes and two edge nodes forming a small-scale cluster that reflects a realistic IoT–Edge/Cloud continuum. IoT (RPi) nodes with cameras connect to the network via Wi-Fi through a 5G CPE device, continuously streaming real-time video to the edge nodes. Edge nodes are deployed in a Multi-access Edge Computing (MEC) environment to emulate near-edge processing capabilities. Telemetry includes system-level metrics such as CPU utilization, memory usage, disk I/O, and network throughput, collected periodically from each node. The dataset is designed to support research in task offloading, AI-driven resource management, and intelligent workload placement across distributed IoT–Edge–Cloud infrastructures. It is particularly useful for developing and validating reinforcement learning (RL) or other data-driven approaches to optimize performance, latency, and energy efficiency in heterogeneous networked systems. PromQL Query CPU Utilization (%): 100 * (1 - avg by (instance) (irate(node_cpu_seconds_total{job="node-exporter", mode=~"idle|iowait|steal", instance="<IP>:9100"}[1m]))) Memory Load (%): 100 * (1 - (node_memory_MemAvailable_bytes{instance="<IP>:9100"} / node_memory_MemTotal_bytes{instance="<IP>:9100"})) Disk I/O Load (bytes/s): sum by (instance) (irate(node_disk_reads_completed_total{instance="<IP>:9100"}[1m]) + irate(node_disk_writes_completed_total{instance="<IP>:9100"}[1m])) Network Download Speed (bytes/s): sum by (instance) (irate(node_network_receive_bytes_total{instance="<IP>:9100", device!~"lo"}[1m])) Network Upload Speed (bytes/s): sum by (instance) (irate(node_network_transmit_bytes_total{instance="<IP>:9100", device!~"lo"}[1m])) Key Features Four-node Kubernetes cluster (2 IoT, 2 Edge) Prometheus-based metric collection with Node Exporter 5G-enabled IoT connectivity via CPE and Wi-Fi Multi-dimensional system telemetry (CPU, memory, disk, network) Time-aligned sampling suitable for machine learning and reinforcement learning Stress and non-stress experimental conditions for controlled workload analysis Potential Use Cases Reinforcement learning for dynamic task offloading Edge–cloud collaboration optimization AI-driven workload orchestration in MEC environments Energy-aware computing in IoT–Edge systems Benchmarking of telemetry-driven scheduling algorithms Format CSV files containing timestamped metric values per node Column structure: [time slot, <node>_cpu, <node>_memory, <node>_disk_io, <node>_net_rx, <node>_net_tx]