feat(sdk): Add DurableInstrumentationPlugin interface and plugin runner

Author: ayushiahjolia

sdk/src/main/java/software/amazon/lambda/durable/DurableConfig.java

@@ -5,6 +5,7 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.time.Duration;
+import java.util.List;
 import java.util.Objects;
 import java.util.Properties;
 import java.util.concurrent.ExecutorService;
@@ -21,6 +22,8 @@
 import software.amazon.lambda.durable.client.DurableExecutionClient;
 import software.amazon.lambda.durable.client.LambdaDurableFunctionsClient;
 import software.amazon.lambda.durable.logging.LoggerConfig;
+import software.amazon.lambda.durable.plugin.DurableExecutionPlugin;
+import software.amazon.lambda.durable.plugin.PluginRunner;
 import software.amazon.lambda.durable.retry.PollingStrategies;
 import software.amazon.lambda.durable.retry.PollingStrategy;
 import software.amazon.lambda.durable.serde.JacksonSerDes;
@@ -94,6 +97,7 @@ public final class DurableConfig {
     private final LoggerConfig loggerConfig;
     private final PollingStrategy pollingStrategy;
     private final Duration checkpointDelay;
+    private final PluginRunner pluginRunner;
 
     private DurableConfig(Builder builder) {
         this.durableExecutionClient = Objects.requireNonNullElseGet(
@@ -104,6 +108,7 @@ private DurableConfig(Builder builder) {
         this.loggerConfig = Objects.requireNonNullElseGet(builder.loggerConfig, LoggerConfig::defaults);
         this.pollingStrategy = Objects.requireNonNullElse(builder.pollingStrategy, PollingStrategies.Presets.DEFAULT);
         this.checkpointDelay = Objects.requireNonNullElseGet(builder.checkpointDelay, () -> Duration.ofSeconds(0));
+        this.pluginRunner = builder.plugins != null ? new PluginRunner(builder.plugins) : PluginRunner.noOp();
 
         validateConfiguration();
     }
@@ -180,6 +185,16 @@ public Duration getCheckpointDelay() {
         return checkpointDelay;
     }
 
+    /**
+     * Gets the plugin runner that dispatches lifecycle events to registered plugins.
+     *
+     * @return PluginRunner instance (never null — returns no-op runner if no plugins configured)
+     * @experimental This method is experimental and may be changed or removed in future releases.
+     */
+    public PluginRunner getPluginRunner() {
+        return pluginRunner;
+    }
+
     public void validateConfiguration() {
         if (getDurableExecutionClient() == null) {
             throw new IllegalStateException("DurableExecutionClient configuration failed");
@@ -274,6 +289,7 @@ public static final class Builder {
         private LoggerConfig loggerConfig;
         private PollingStrategy pollingStrategy;
         private Duration checkpointDelay;
+        private List<DurableExecutionPlugin> plugins;
 
         public Builder() {}
 
@@ -383,6 +399,32 @@ public Builder withCheckpointDelay(Duration duration) {
             return this;
         }
 
+        /**
+         * Sets instrumentation plugins for observability and tracing.
+         *
+         * <p>Plugins receive lifecycle callbacks at key points during durable execution, enabling integration with
+         * tracing systems (e.g., OpenTelemetry, X-Ray), custom metrics, and logging enrichment.
+         *
+         * <p>Multiple plugins can be provided and will be called in order. Plugin errors are swallowed to prevent
+         * instrumentation from affecting execution correctness.
+         *
+         * <p>Example:
+         *
+         * <pre>{@code
+         * DurableConfig.builder()
+         *     .withPlugins(List.of(new OpenTelemetryDurablePlugin()))
+         *     .build();
+         * }</pre>
+         *
+         * @param plugins list of instrumentation plugins
+         * @return This builder
+         * @experimental This method is experimental and may be changed or removed in future releases.
+         */
+        public Builder withPlugins(List<DurableExecutionPlugin> plugins) {
+            this.plugins = plugins;
+            return this;
+        }
+
         /**
          * Builds the DurableConfig instance.
          *

sdk/src/main/java/software/amazon/lambda/durable/plugin/DurableExecutionPlugin.java

@@ -0,0 +1,71 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+/**
+ * Plugin interface for instrumenting durable execution lifecycle events.
+ *
+ * <p>Implement this interface to integrate observability tools (OpenTelemetry, Datadog, etc.) with the durable
+ * execution SDK. The SDK calls these hooks at key lifecycle points without requiring modifications to core SDK code.
+ *
+ * <p>All methods have default no-op implementations, allowing plugins to override only the hooks they need.
+ *
+ * <p>Plugin errors are isolated — exceptions thrown by plugin methods are caught and logged but never disrupt SDK
+ * execution.
+ *
+ * <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public interface DurableExecutionPlugin {
+
+    // ─── Invocation-level hooks ──────────────────────────────────────────
+
+    /**
+     * Called at the start of each Lambda invocation. Use to set up per-invocation state (trace ID, invocation span).
+     *
+     * <p>Check {@link InvocationInfo#isFirstInvocation()} to detect the first invocation of an execution (useful for
+     * sampling decisions or execution-level span creation).
+     */
+    default void onInvocationStart(InvocationInfo info) {}
+
+    /**
+     * Called at the end of each Lambda invocation. Use to flush spans/metrics before Lambda freezes.
+     *
+     * <p>This hook is awaited — the SDK blocks until it returns. This is the only safe flush point before Lambda
+     * freezes the execution environment.
+     *
+     * <p>Check {@link InvocationEndInfo#invocationStatus()} to detect if the execution reached a terminal state in this
+     * invocation (useful for writing summary records or flushing final data).
+     */
+    default void onInvocationEnd(InvocationEndInfo info) {}
+
+    // ─── Operation-level hooks ───────────────────────────────────────────
+
+    /** Called when an operation starts (including replay). Use for logging/metrics that want replay visibility. */
+    default void onOperationStart(OperationInfo info) {}
+
+    /**
+     * Called when an operation reaches a terminal status for the first time (not on replay).
+     *
+     * <p>The OTel plugin creates operation spans here with backfilled start/end timestamps.
+     */
+    default void onOperationEnd(OperationEndInfo info) {}
+
+    // ─── User function hooks ─────────────────────────────────────────────
+
+    /**
+     * Called when a user-provided function starts executing. This fires for both step attempts (with {@code attempt}
+     * set) and child context functions (with {@code attempt} null).
+     *
+     * <p>This hook fires on the same thread as user code, so plugins can set OTel context via
+     * {@code Context.makeCurrent()} here.
+     */
+    default void onUserFunctionStart(UserFunctionStartInfo info) {}
+
+    /**
+     * Called when a user-provided function finishes executing. This fires for both step attempts and child context
+     * functions.
+     *
+     * <p>This hook fires on the same thread as user code, so plugins can close OTel scopes here.
+     */
+    default void onUserFunctionEnd(UserFunctionEndInfo info) {}
+}

sdk/src/main/java/software/amazon/lambda/durable/plugin/InvocationEndInfo.java

@@ -0,0 +1,20 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+/**
+ * Information provided at the end of a Lambda invocation.
+ *
+ * @param requestId the Lambda request ID for this invocation
+ * @param executionArn the durable execution ARN
+ * @param isFirstInvocation true if this is the first invocation of the execution
+ * @param invocationStatus the invocation outcome (SUCCEEDED, FAILED, or PENDING)
+ * @param executionError non-null if the execution failed
+ *     <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public record InvocationEndInfo(
+        String requestId,
+        String executionArn,
+        boolean isFirstInvocation,
+        InvocationStatus invocationStatus,
+        Throwable executionError) {}

sdk/src/main/java/software/amazon/lambda/durable/plugin/InvocationInfo.java

@@ -0,0 +1,13 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+/**
+ * Invocation-level information available to plugin hooks.
+ *
+ * @param requestId the Lambda request ID for this invocation
+ * @param executionArn the durable execution ARN
+ * @param isFirstInvocation true if this is the first invocation of the execution (not a replay invocation)
+ *     <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public record InvocationInfo(String requestId, String executionArn, boolean isFirstInvocation) {}

sdk/src/main/java/software/amazon/lambda/durable/plugin/InvocationStatus.java

@@ -0,0 +1,17 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+/**
+ * Status of a Lambda invocation at the end of its execution.
+ *
+ * <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public enum InvocationStatus {
+    /** Execution completed successfully in this invocation. */
+    SUCCEEDED,
+    /** Execution failed in this invocation. */
+    FAILED,
+    /** Execution suspended — will resume in a future invocation. */
+    PENDING
+}

sdk/src/main/java/software/amazon/lambda/durable/plugin/OperationEndInfo.java

@@ -0,0 +1,32 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+import java.time.Instant;
+
+/**
+ * Extended operation information for operation end events.
+ *
+ * @param id operation ID — hashed
+ * @param rawId operation ID — unhashed
+ * @param name human-readable operation name (may be null)
+ * @param type operation type
+ * @param subType operation sub-type (may be null)
+ * @param parentId parent operation ID — hashed (null for root-level operations)
+ * @param rawParentId parent operation ID — unhashed (null for root-level operations)
+ * @param startTimestamp when the operation started
+ * @param endTimestamp when the operation ended
+ * @param error non-null if the operation failed
+ *     <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public record OperationEndInfo(
+        String id,
+        String rawId,
+        String name,
+        String type,
+        String subType,
+        String parentId,
+        String rawParentId,
+        Instant startTimestamp,
+        Instant endTimestamp,
+        Throwable error) {}

sdk/src/main/java/software/amazon/lambda/durable/plugin/OperationInfo.java

@@ -0,0 +1,32 @@
+// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+// SPDX-License-Identifier: Apache-2.0
+package software.amazon.lambda.durable.plugin;
+
+import java.time.Instant;
+
+/**
+ * Operation-level information available to plugin hooks.
+ *
+ * <p>Field names mirror the {@code Operation} type from the AWS SDK for consistency.
+ *
+ * @param id operation ID — hashed (unique within the execution)
+ * @param rawId operation ID — unhashed (e.g. "1", "2", "hash(1)-1" for child contexts)
+ * @param name human-readable operation name (may be null)
+ * @param type operation type (STEP, WAIT, CONTEXT, CHAINED_INVOKE, CALLBACK)
+ * @param subType operation sub-type (Map, Parallel, RunInChildContext, WaitForCondition, etc.) — may be null
+ * @param parentId parent operation ID — hashed (null for root-level operations)
+ * @param rawParentId parent operation ID — unhashed (null for root-level operations)
+ * @param startTimestamp when the operation started (may be from a prior invocation)
+ * @param endTimestamp when the operation ended (null if still running)
+ *     <p><b>Preview API:</b> This API is experimental and may be changed or removed in future releases.
+ */
+public record OperationInfo(
+        String id,
+        String rawId,
+        String name,
+        String type,
+        String subType,
+        String parentId,
+        String rawParentId,
+        Instant startTimestamp,
+        Instant endTimestamp) {}