Monitoring
Fogify provides a wide rage of monitoring metrics for each emulated instance and is able to extract user-defined metrics.
Default Metrics
Fogify captures performance metrics directly from the containerized process by inspecting the containers.
Pre-defined metrics available by Fogify include CPU time, memory usage, network traffic and many more.
The following table depicts all available metrics provided by Fogify
| metric | description |
|---|---|
| cpu | The accumulative CPU time occupied by an emulated instance |
| cpu_util | The average utilization of CPU between two measurements |
| memory | The occupied memory by an emulated instance |
| memory_util | The utilization of the emulated instance’s memory |
| network_rx_{network name} | The accumulative received bytes from a specific network |
| network_tx_{network name} | The accumulative transmitted bytes from a specific network |
User-defined Metrics
Users can expose application-level metrics for their IoT services.
This is achieved by exposing metric updates, in JSON format, to a file (fogify.metrics.json) via the container root directory interface.
Fogify will then retrieve them via a lightweight monitoring probe.
When the agent retrieves the data, it stores them to the monitoring storage.
Following, we offer an exemplary fogify.metrics.json file.
{
"response-time": 500,
"requests": 6530,
"log-in-requests": 100
}
In the previous json file there are three metrics exposed by the application, namely, the response time (response-time),
requests count (requests) and login request count (log-in-requests).