Partager via


Microsoft.SystemInsights.Capability Namespace

Classes

AnomalyDetectionDataPoint

This class defines a data point specific to anomaly detection predictions, where each data point is associated with an anomaly label.

AnomalyDetectionResult

This class represents the result of an anomaly detection prediction.

CapabilityInformation

This class represents a capability and associated information, including the capability description, prediction type, and data sources,

DataRecord

This class contains all of the data associated with a specific timestamp. For certain data sources, multiple data points are generated for each timestamp. For example, the network well-known series reports both BytesInbound, BytesOutbound, and BytesTotal for each timestamp.

The order of data points within each DataRecord is specified in the SeriesNames dictionary.

For example, the SeriesNames object for the volume well-known series volume looks like: { {Size, 0} {SizeUsed, 1} {SizeFree, 2} } This means that for a given timestamp DateTime, there exists an array of three elements (Values), which looks like [{SizeValue}, {SizeUsedValue}, {SizeFreeValue}].

For the CPU well-known series, the SeriesNames looks like: { {ProcessorTime, 0} } So, for a given timestamp DateTime, there's an array with one value Values, which looks like [{ProcessorTimeValue}].

For the networking well-known series, SeriesNames looks like:
{ {BytesMaximum, 0} {BytesInbound, 1} {BytesOutbound, 2} {BytesTotal, 3} } So, for a given timestamp DateTime, the Values is an array of 4 elements, which looks like [{BytesMaximumValue}, {BytesInboundValue}, {BytesOutboundValue}, {BytesTotal}].

For the physical disk well-known series, SeriesNames looks like:
{ {Size, 0} } So, for a given timestamp DateTime, the Values is an array of 1 element, which looks like [{SizeValue}].

For each instance of a performance counter,SeriesNames looks like: { {"CounterValue", 0} } So, for a given timestamp DateTime, the Values is an array of 1 element, which looks like [{CounterValue}].

For each event data source, the SeriesNames looks like: { {"EventCount", 0} } So, for a given timestamp DateTime, the Values is an array of 1 element, which looks like [{EventCount}].

DataSeries

This class contains the data associated with a data source. Each data source that's registered for a capability maps directly to a DataSeries object.

An individual data source might record data for multiple instances. For example, the well-known volume series has data for all volumes. Or, a performance counter data source can record data for 'all instances'. The DataSeries object contains structures to help you properly identify the data you specified.

Identifier helps identify the performance counter instance or the logical object, such as an individual volume or network adapter.

SeriesProperties contains any additional metadata associated with the data source, such as a GUID or FriendlyName.

SeriesNames specifies the ordering of the data contained in each DataRecord object.

DataRecords contains a DataRecord for each timestamp.

For example, in the volume well-known series DataSeries, Identifier: VolumeGuid SeriesProperties: { { "VolumeGuid", value } { "FriendlyName", value } { "DriveLetter", value } } SeriesName: { { Size, 0 } { SizeUsed, 1 } { SizeFree, 2 } }

For example, in the CPU well-known series DataSeries, Identifier: "CPU" SeriesProperties: { { "Id", value } } SeriesNames: { { ProcessorTime, 0 } }

For example, in the networking well-known series DataSeries, Identifier: AdapterGuid SeriesProperties: { { "AdapterGuid", value } { "FriendlyName", value } } SeriesNames: { {BytesMaximum, 0} {BytesInbound, 1} {BytesOutbound, 2} {BytesTotal, 3} }

For example, in the physical disk well-known series DataSeries, Identifier: Id SeriesProperties: { { "Id", value } } SeriesNames: { { Size, 0} }

For example, for the performance counter DataSeries, Identifier: Counter Instance Name SeriesProperties: { { "Path", value } { "Name", value } { "Instance", value } } SeriesNames: { {"CounterValue", 0} }

For example, in the event DataSeries, Identifier: EventId, SeriesProperties: { { "ChannelName", value } { "EventId", value } } SeriesNames: { {"EventCount", 0} }

DataSource

This class represents the data collected by System Insights.

EventDataSource

This class represents an event log data source.

ForecastingResult

The ForecastingResult class represents the result of a forecasting prediction.

InvokeRequest

This class represents the object being passed to the capability when it is invoked.

InvokeResult

This class represents the result object returned by each capability after it's invoked.

Each capability may forecast or analyze multiple series, generating multiple results. This class allows you to output the result of each of those series, and then specify a higher level status and a status description to describe the overall result of the prediction. Only one of ForecastingResults, AnomalyDetectionResults, GenericResult will be valid

PerformanceCounterDataSource

This class represents a performance counter data source.

WellKnownSeriesDataSource

This class represents a well-known series data source.

Interfaces

ICapability

Defines the interface that each System Insights capability must implement.

Enums

AnomalyType

This enum specifies different severities of anomalies.