Analyzers

Analyzers perform statistical computations on your observability data — forecasting, change-point detection, correlation, and anomaly detection. You list the available analyzers, inspect an analyzer’s input/output schema, run one against a DQL timeseries, and validate input before executing.

List Analyzers

# List all available analyzers
dtctl get analyzers

# Get a single analyzer's raw definition
dtctl get analyzer dt.statistics.GenericForecastAnalyzer

# Filter the list (DQL-style expression)
dtctl get analyzers --filter "name contains 'forecast'"

# Structured output
dtctl get analyzers -o json

Describe an Analyzer

describe resolves an analyzer’s JSON Schemas and renders the required and optional input fields so you know exactly what to pass to exec analyzer. Unlike get analyzer, which returns the raw definition, describe flattens the input and result schemas into a readable table (and includes them in JSON/YAML output).

# Describe an analyzer and see its input fields
dtctl describe analyzer dt.statistics.GenericForecastAnalyzer

# Print the analyzer's full markdown documentation
dtctl describe analyzer dt.statistics.GenericForecastAnalyzer --doc

# Structured output (includes inputSchema and resultSchema)
dtctl describe analyzer dt.statistics.GenericForecastAnalyzer -o json

Example output:

Name:         dt.statistics.GenericForecastAnalyzer
Display Name: Forecast Analysis
Category:     Forecasting
Type:         AnalyzerFactory

Input (required):
  timeSeriesData   string    DQL timeseries query to forecast
  forecastHorizon  integer   Number of intervals to predict

Input (optional):
  coverageProbability  number   Confidence band width (e.g. 0.9)

Output:
  forecastQualityAssessment  object   Quality metrics for the forecast
  output                     array    Predicted values with confidence bands

  Run it:  dtctl exec analyzer dt.statistics.GenericForecastAnalyzer --query <dql>
  Docs:    dtctl describe analyzer dt.statistics.GenericForecastAnalyzer --doc

analyzer, analyzers, and az are interchangeable aliases across get, describe, exec, and verify.

Execute an Analyzer

# Run a forecast analyzer with the DQL query shorthand
# (works for timeseries-based analyzers)
dtctl exec analyzer dt.statistics.GenericForecastAnalyzer \
  --query "timeseries avg(dt.host.cpu.usage)"

# Provide the full input as inline JSON
dtctl exec analyzer dt.statistics.GenericForecastAnalyzer \
  --input '{"query":"timeseries avg(dt.host.cpu.usage)"}'

# Provide input from a JSON file
dtctl exec analyzer dt.statistics.GenericForecastAnalyzer -f input.json

# Execute and wait for completion (default; --wait=false returns immediately)
dtctl exec analyzer dt.statistics.GenericForecastAnalyzer \
  -f input.json --wait --timeout 300

# Structured output
dtctl exec analyzer dt.statistics.GenericForecastAnalyzer -f input.json -o json

Flag	Description
`--query`	DQL timeseries query shorthand (for timeseries analyzers)
`--input`	Inline JSON input
`-f`, `--file`	Read JSON input from a file
`--validate`	Validate input without executing
`--wait`	Wait for execution to complete (default `true`)
`--timeout`	Timeout in seconds when waiting (default `300`)

Validate Input

Validate an analyzer input against its schema without running it. This is the same check exec analyzer --validate performs, exposed as its own verb.

# Validate input from a file
dtctl verify analyzer dt.statistics.GenericForecastAnalyzer -f input.json

# Validate inline JSON
dtctl verify analyzer dt.statistics.GenericForecastAnalyzer \
  --input '{"timeSeriesData":"timeseries avg(dt.host.cpu.usage)"}'

# Validate the DQL query shorthand
dtctl verify analyzer dt.statistics.GenericForecastAnalyzer \
  --query "timeseries avg(dt.host.cpu.usage)"

# Structured output
dtctl verify analyzer dt.statistics.GenericForecastAnalyzer -f input.json -o json

Common Analyzers

Analyzer	Description
`dt.statistics.GenericForecastAnalyzer`	Predict future metric values based on historical trends
`dt.statistics.GenericChangePointAnalyzer`	Detect significant changes in metric behavior
`dt.statistics.GenericCorrelationAnalyzer`	Find correlations between metric time series
`dt.statistics.GenericAnomalyDetectionAnalyzer`	Identify anomalous metric patterns

Use dtctl get analyzers to discover every analyzer available in your environment, then dtctl describe analyzer <name> to see its inputs.

Required Scopes

Scope	Used By
`davis:analyzers:read`	Listing and describing analyzers
`davis:analyzers:execute`	Executing and validating analyzers