Provide capability to configure the metric aggregation interval per metric

[tomkerkhove]

Metrics do not always need to be scraped with the same metrics aggregation.

We should provide the capability to configure the metric aggregation interval per metric.

Checklist

In order to achieve this we will:

• Introduce aggregation as part of azureMetricConfiguration which contains all aggregation related information
• Configure the aggregation interval via azureMetricConfiguration.aggregation.interval
• Move aggregation type ot the aggregation interval via azureMetricConfiguration.aggregation.type
• Use configured aggregation period when scraping Azure Monitor
• Update documentation on how to configure this and the impact of the size
• Provide clear guidance on breaking change in documentation (depends on Introduce a change log on promitor.io #255)

What needs to be defined

• How we will configure our aggragation interval - Do we stick with cron or do we use a different approach?
• Do we enforce intervals higher than 5 minutes and validate it at startup, or is it the responsibility of the consumer (assuming we make this clear in docs)

Metrics Specification

azureMetadata:
  tenantId: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
  subscriptionId: yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy
  resourceGroupName: promitor
metrics: 
  - name: demo_queue_size
    description: "Amount of active messages of the 'myqueue' queue"
    resourceType: ServiceBusQueue
    namespace: promitor-messaging
    queueName: orders
    azureMetricConfiguration:
      metricName: ActiveMessages
      aggregation:
        type: Total
        interval: <interval>

Other

Closes #60 and #198

[tomkerkhove]

Will pick this up next week probably

[tomkerkhove]

Docs will be handled via #293