MDA Version 1.3

Updates

1. Aggregation Operations

The previous version (1.2) disabled the following functionality: monitor a metric that needs to be aggregated within a certain timestamp. This version incorporates this operation. With the updated version, the monitoring spec, in case that exists an aggregable metric, needs to include two extra fields, as displayed here:

"metrics": [
        {
            "metricName": "cpu_utilization",
            "metricType": "float",
            "aggregationMethod": "sum",
            "step": "1m",
            "step_aggregation": "5m",
            "next_run_at": null
        }
    ]

Concerning these fields, the "aggregationMethod" needs to include a valid operation. For now, we make available the methods: sum, average, count, max, min, and stddev. For each metric, there is a 1. step that defines the interval that it is fetched values from a resource; 2. aggregation step that sets the timeline where the values are aggregable and pushed to the Data Lake.