Class SpringAIObservabilityHandler

java.lang.Object

com.google.adk.models.springai.observability.SpringAIObservabilityHandler

public class SpringAIObservabilityHandler
extends Object

Handles observability features for Spring AI integration using Micrometer.

This class provides:

• Metrics collection for request latency, token counts, and error rates via Micrometer
• Request/response logging with configurable content inclusion
• Performance monitoring for streaming and non-streaming requests
• Integration with any Micrometer-compatible metrics backend (Prometheus, Datadog, etc.)

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	SpringAIObservabilityHandler.RequestContext	Context for tracking a single request with Micrometer timer.

Constructor Summary

Constructors

Constructor	Description
SpringAIObservabilityHandler(SpringAIProperties.Observability config)	Creates an observability handler with a default SimpleMeterRegistry.
SpringAIObservabilityHandler(SpringAIProperties.Observability config, io.micrometer.core.instrument.MeterRegistry meterRegistry)	Creates an observability handler with a custom MeterRegistry.

Method Summary

Modifier and Type	Method	Description
io.micrometer.core.instrument.MeterRegistry	getMeterRegistry()	Gets the Micrometer MeterRegistry for direct access to metrics.
void	logRequest(String content, String modelName)	Logs request content if enabled.
void	logResponse(String content, String modelName)	Logs response content if enabled.
void	recordError(SpringAIObservabilityHandler.RequestContext context, Throwable error)	Records a failed request.
void	recordSuccess(SpringAIObservabilityHandler.RequestContext context, int tokenCount, int inputTokens, int outputTokens)	Records the completion of a successful request.
SpringAIObservabilityHandler.RequestContext	startRequest(String modelName, String requestType)	Records the start of a request.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

SpringAIObservabilityHandler

public SpringAIObservabilityHandler(SpringAIProperties.Observability config)

Creates an observability handler with a default SimpleMeterRegistry.

Parameters:

config - the observability configuration

SpringAIObservabilityHandler

public SpringAIObservabilityHandler(SpringAIProperties.Observability config,
 io.micrometer.core.instrument.MeterRegistry meterRegistry)

Creates an observability handler with a custom MeterRegistry.

Parameters:

config - the observability configuration

meterRegistry - the Micrometer meter registry to use for metrics

Method Details

startRequest

public SpringAIObservabilityHandler.RequestContext startRequest(String modelName,
 String requestType)

Records the start of a request.

Parameters:

modelName - the name of the model being used

requestType - the type of request (e.g., "chat", "streaming")

Returns:

a request context for tracking the request

recordSuccess

public void recordSuccess(SpringAIObservabilityHandler.RequestContext context,
 int tokenCount,
 int inputTokens,
 int outputTokens)

Records the completion of a successful request.

Parameters:

context - the request context

tokenCount - the number of tokens processed (input + output)

inputTokens - the number of input tokens

outputTokens - the number of output tokens

recordError

public void recordError(SpringAIObservabilityHandler.RequestContext context,
 Throwable error)

Records a failed request.

Parameters:

context - the request context

error - the error that occurred

logRequest

public void logRequest(String content,
 String modelName)

Logs request content if enabled.

Parameters:

content - the request content

modelName - the model name

logResponse

public void logResponse(String content,
 String modelName)

Logs response content if enabled.

Parameters:

content - the response content

modelName - the model name

getMeterRegistry

public io.micrometer.core.instrument.MeterRegistry getMeterRegistry()

Gets the Micrometer MeterRegistry for direct access to metrics.

This allows users to export metrics to any Micrometer-compatible backend (Prometheus, Datadog, CloudWatch, etc.) or query metrics programmatically.

Returns:

the MeterRegistry instance