Skip to main content

MetricTimeSeries

Introduction​

Metric Timeseries represent a timeseries of metrics from various metric providers.

A metric timeseries consists of the following:

Provider​

The name of the metric provider

Metric​

The name of the metric to read from the provider

info

See Metrics for full details on providers and metric names

Filter​

An optional filter to filter the metrics in the value calculation

Calendar​

This is the interval frequencies for value aggregation, usually one of:

  • #PT1M - 1 minutely
  • #PT5M - 5 minutely
  • #PT15M - 15 minutely
  • #PT30M - 30 minutely
  • #PT1H - hourly
  • #PT6H - 6 hourly
  • #PT12H - 12 hourly
  • #P1D - daily

Observed​

The aggregation method, usually one of:

  • summed - Sum up all the values for the interval
  • averaged - Average the values for the interval
  • high - The maximum value during the interval
  • low - The minimum value during the interval

Creating a MetricTimeSeries​

Azure Service Bus Metrics​

An example of creating a metric timeseries for the average number of active messages on a queue named etrm aggregated 5 minutely

ActiveMessages_1h = MetricTimeSeries("AzureServiceBus", "ActiveMessages")
ActiveMessages_1h.calendar = ${calendar:"#PT5M"}
ActiveMessages_1h.queue = "etrm"
ActiveMessages_1h.observed="averaged"

ODSL User Request Metrics​

An example of creating metric timeseries for the total user read requests for the service 'user' aggregated hourly 

ReadRequests_1h = MetricTimeSeries("ODSLRequest", "ReadRequests")
ReadRequests_1h.calendar = ${calendar:"#PT1H"}
ReadRequests_1h.observed = "summed"
ReadRequests_1h.filter = ?service='user'