Lambda Deep Dive Basics

Q: What status code should a synchronous invoke return?

200. An asynchronous (Event) invoke returns 202, and a DryRun returns 204. The status reflects the invoke itself, not whether your handler logic succeeded.

Q: How do I tell if my handler threw an error?

Check for FunctionError in the response. If present (Unhandled or Handled), the handler failed even though the HTTP call returned 200. The payload then holds the error object.

Q: How do I get logs without CloudWatch?

Set LogType: "Tail" on a synchronous invoke to get the last 4 KB in LogResult, base64-encoded. For full history, read the function's CloudWatch Logs group.

Q: Does update_function_configuration replace all settings?

Only the fields you pass, but structured fields like Environment.Variables are replaced wholesale. Merge in the keys you want to keep before sending.

Q: Can I update code and publish a version at once?

Yes. Pass Publish: true to update_function_code, and Lambda publishes a new immutable version from the new code without a separate PublishVersion call.

Q: Why do I get ResourceConflictException on back-to-back updates?

The previous change is still applying. Wait with the function_updated_v2 waiter until LastUpdateStatus is Successful, then issue the next update.

Six examples to get comfortable managing and running a Lambda function from an SDK - four basic and two intermediate.

Each example shows the same operation in both Python (boto3) and TypeScript (AWS SDK for JavaScript v3). By the end you can push new code, retune a function's resources, and invoke it both ways while reading back its result and logs.

Prerequisites

Python: pip install boto3 (boto3 1.43.x, Python 3.10+). TypeScript: npm install @aws-sdk/client-lambda (Node.js 18+).
Credentials configured via aws configure, environment variables, or an IAM role - the SDK finds them automatically.
An existing function named my-func and IAM permissions for lambda:UpdateFunctionCode, lambda:UpdateFunctionConfiguration, lambda:GetFunctionConfiguration, and lambda:InvokeFunction.

Basic Examples

1. Invoke Synchronously and Read the Result

A synchronous (RequestResponse) invoke runs the function and returns its output on the same call.

# --- Python (boto3) ---
import boto3, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="RequestResponse",
    Payload=json.dumps({"name": "Ada"}).encode(),
)
result = json.loads(resp["Payload"].read())
print(resp["StatusCode"], result)

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "RequestResponse",
  Payload: new TextEncoder().encode(JSON.stringify({ name: "Ada" })),
}));
const result = JSON.parse(new TextDecoder().decode(resp.Payload));
console.log(resp.StatusCode, result);

StatusCode is 200 for a completed synchronous invoke.
Payload is a streaming body in boto3 (.read()) and a Uint8Array in SDK v3 (TextDecoder).
The payload is whatever your handler returns, serialized as JSON.
Always decode before json.loads / JSON.parse - the SDK gives you bytes, not a string.

Related: Lambda's Execution Model for SDK Callers - why synchronous invokes pay cold-start latency inline.

2. Invoke Asynchronously

An asynchronous (Event) invoke queues the payload and returns immediately, letting Lambda run and retry it for you.

# --- Python (boto3) ---
import boto3, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="Event",
    Payload=json.dumps({"job": "resize", "id": 42}).encode(),
)
print(resp["StatusCode"])  # 202 - accepted, not yet run

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "Event",
  Payload: new TextEncoder().encode(JSON.stringify({ job: "resize", id: 42 })),
}));
console.log(resp.StatusCode); // 202 - accepted, not yet run

StatusCode is 202: Lambda accepted the event but has not run it yet.
There is no return payload - the response body is empty by design.
Lambda retries a failed async invoke twice, then drops it or sends it to a configured destination.
Use this for fire-and-forget work where the caller should not wait.

3. Update Function Configuration

Change memory, timeout, and environment variables without touching the code.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
lam.update_function_configuration(
    FunctionName="my-func",
    MemorySize=512,          # MB, 128-10240
    Timeout=30,              # seconds, max 900
    Environment={"Variables": {"LOG_LEVEL": "info"}},
)

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, UpdateFunctionConfigurationCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionConfigurationCommand({
  FunctionName: "my-func",
  MemorySize: 512,          // MB, 128-10240
  Timeout: 30,              // seconds, max 900
  Environment: { Variables: { LOG_LEVEL: "info" } },
}));

More MemorySize also grants proportionally more CPU, so it can speed up compute-bound functions.
Environment.Variables replaces the whole map - include existing keys you want to keep.
The call returns while the change applies; the function is briefly LastUpdateStatus: InProgress.
Timeout caps a single invocation; the hard maximum is 900 seconds (15 minutes).

4. Update Function Code

Ship a new deployment package from a local ZIP.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
with open("function.zip", "rb") as f:
    lam.update_function_code(
        FunctionName="my-func",
        ZipFile=f.read(),
        Publish=True,   # also cut an immutable version
    )

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, UpdateFunctionCodeCommand } from "@aws-sdk/client-lambda";
import { readFileSync } from "node:fs";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionCodeCommand({
  FunctionName: "my-func",
  ZipFile: readFileSync("function.zip"),
  Publish: true,   // also cut an immutable version
}));

ZipFile takes the raw bytes of the package; for large artifacts pass S3Bucket/S3Key instead.
Publish: true publishes a new numbered version from the updated code in one step.
For container-image functions, pass ImageUri rather than ZipFile.
As with config changes, the function goes Pending while the new code is deployed.

Related: Versions, Aliases & Traffic Shifting - what Publish: true gives you and how to route to it.

Intermediate Examples

5. Capture Tail Logs and Detect a Handler Error

Ask Lambda to return the last chunk of logs and check whether the function itself failed.

# --- Python (boto3) ---
import boto3, base64, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="RequestResponse",
    LogType="Tail",   # include last 4 KB of logs
    Payload=json.dumps({"name": "Ada"}).encode(),
)
if "FunctionError" in resp:
    print("handler failed:", resp["FunctionError"])
print(base64.b64decode(resp["LogResult"]).decode())

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "RequestResponse",
  LogType: "Tail",   // include last 4 KB of logs
  Payload: new TextEncoder().encode(JSON.stringify({ name: "Ada" })),
}));
if (resp.FunctionError) console.log("handler failed:", resp.FunctionError);
console.log(Buffer.from(resp.LogResult ?? "", "base64").toString());

LogType: "Tail" returns LogResult, base64-encoded, holding the last 4 KB of logs.
FunctionError is set (Unhandled or Handled) when the handler throws - even though the HTTP call itself succeeded.
A 200 status with a FunctionError is a successful invoke of a failing function - check both.
LogType only applies to synchronous invokes.

6. Wait Until an Update Finishes

Config and code updates are eventually consistent; wait for Active/Successful before the next step.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
lam.update_function_configuration(FunctionName="my-func", MemorySize=1024)
 
waiter = lam.get_waiter("function_updated_v2")
waiter.wait(FunctionName="my-func")   # returns when LastUpdateStatus is Successful
print("update applied")

// --- TypeScript (AWS SDK v3) ---
import {
  LambdaClient,
  UpdateFunctionConfigurationCommand,
  waitUntilFunctionUpdatedV2,
} from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionConfigurationCommand({ FunctionName: "my-func", MemorySize: 1024 }));
 
const r = await waitUntilFunctionUpdatedV2({ client: lam, maxWaitTime: 120 }, { FunctionName: "my-func" });
if (r.state !== "SUCCESS") throw new Error(`update not applied: ${r.state}`);
console.log("update applied");

The function_updated_v2 waiter polls GetFunctionConfiguration until LastUpdateStatus is Successful.
Chaining updates without waiting risks ResourceConflictException.
SDK v3's waitUntilFunctionUpdatedV2 returns a state object; branch on SUCCESS instead of catching.
There is a matching function_active_v2 / waitUntilFunctionActiveV2 for newly created functions.

FAQs

What status code should a synchronous invoke return?

200. An asynchronous (Event) invoke returns 202, and a DryRun returns 204. The status reflects the invoke itself, not whether your handler logic succeeded.

How do I tell if my handler threw an error?

Check for FunctionError in the response. If present (Unhandled or Handled), the handler failed even though the HTTP call returned 200. The payload then holds the error object.

Why is Payload bytes and not a string?

The SDKs return the raw response body. boto3 gives a streaming body you .read(); SDK v3 gives a Uint8Array you decode with TextDecoder. Then json.loads / JSON.parse it.

How do I get logs without CloudWatch?

Set LogType: "Tail" on a synchronous invoke to get the last 4 KB in LogResult, base64-encoded. For full history, read the function's CloudWatch Logs group.

Does update_function_configuration replace all settings?

Only the fields you pass, but structured fields like Environment.Variables are replaced wholesale. Merge in the keys you want to keep before sending.

Can I update code and publish a version at once?

Yes. Pass Publish: true to update_function_code, and Lambda publishes a new immutable version from the new code without a separate PublishVersion call.

Why do I get ResourceConflictException on back-to-back updates?

The previous change is still applying. Wait with the function_updated_v2 waiter until LastUpdateStatus is Successful, then issue the next update.

What is the maximum timeout and memory?

Timeout is up to 900 seconds (15 minutes). Memory ranges from 128 MB to 10,240 MB, and CPU scales with it.

Lambda's Execution Model for SDK Callers - the model behind these calls.
Versions, Aliases & Traffic Shifting - publishing and routing after you update code.
Layers & Shared Dependencies - attach shared code and dependencies via configuration.
Lambda via SDK Best Practices - the defaults to set for invokes and updates.

Stack versions: This page was written for boto3 1.43.x (Python 3.10+) and the AWS SDK for JavaScript v3 (Node.js 18+).

Prerequisites

Python: pip install boto3 (boto3 1.43.x, Python 3.10+). TypeScript: npm install @aws-sdk/client-lambda (Node.js 18+).
Credentials configured via aws configure, environment variables, or an IAM role - the SDK finds them automatically.
An existing function named my-func and IAM permissions for lambda:UpdateFunctionCode, lambda:UpdateFunctionConfiguration, lambda:GetFunctionConfiguration, and lambda:InvokeFunction.

Basic Examples

1. Invoke Synchronously and Read the Result

A synchronous (RequestResponse) invoke runs the function and returns its output on the same call.

# --- Python (boto3) ---
import boto3, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="RequestResponse",
    Payload=json.dumps({"name": "Ada"}).encode(),
)
result = json.loads(resp["Payload"].read())
print(resp["StatusCode"], result)

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "RequestResponse",
  Payload: new TextEncoder().encode(JSON.stringify({ name: "Ada" })),
}));
const result = JSON.parse(new TextDecoder().decode(resp.Payload));
console.log(resp.StatusCode, result);

StatusCode is 200 for a completed synchronous invoke.
Payload is a streaming body in boto3 (.read()) and a Uint8Array in SDK v3 (TextDecoder).
The payload is whatever your handler returns, serialized as JSON.
Always decode before json.loads / JSON.parse - the SDK gives you bytes, not a string.

Related: Lambda's Execution Model for SDK Callers - why synchronous invokes pay cold-start latency inline.

2. Invoke Asynchronously

An asynchronous (Event) invoke queues the payload and returns immediately, letting Lambda run and retry it for you.

# --- Python (boto3) ---
import boto3, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="Event",
    Payload=json.dumps({"job": "resize", "id": 42}).encode(),
)
print(resp["StatusCode"])  # 202 - accepted, not yet run

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "Event",
  Payload: new TextEncoder().encode(JSON.stringify({ job: "resize", id: 42 })),
}));
console.log(resp.StatusCode); // 202 - accepted, not yet run

StatusCode is 202: Lambda accepted the event but has not run it yet.
There is no return payload - the response body is empty by design.
Lambda retries a failed async invoke twice, then drops it or sends it to a configured destination.
Use this for fire-and-forget work where the caller should not wait.

3. Update Function Configuration

Change memory, timeout, and environment variables without touching the code.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
lam.update_function_configuration(
    FunctionName="my-func",
    MemorySize=512,          # MB, 128-10240
    Timeout=30,              # seconds, max 900
    Environment={"Variables": {"LOG_LEVEL": "info"}},
)

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, UpdateFunctionConfigurationCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionConfigurationCommand({
  FunctionName: "my-func",
  MemorySize: 512,          // MB, 128-10240
  Timeout: 30,              // seconds, max 900
  Environment: { Variables: { LOG_LEVEL: "info" } },
}));

More MemorySize also grants proportionally more CPU, so it can speed up compute-bound functions.
Environment.Variables replaces the whole map - include existing keys you want to keep.
The call returns while the change applies; the function is briefly LastUpdateStatus: InProgress.
Timeout caps a single invocation; the hard maximum is 900 seconds (15 minutes).

4. Update Function Code

Ship a new deployment package from a local ZIP.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
with open("function.zip", "rb") as f:
    lam.update_function_code(
        FunctionName="my-func",
        ZipFile=f.read(),
        Publish=True,   # also cut an immutable version
    )

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, UpdateFunctionCodeCommand } from "@aws-sdk/client-lambda";
import { readFileSync } from "node:fs";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionCodeCommand({
  FunctionName: "my-func",
  ZipFile: readFileSync("function.zip"),
  Publish: true,   // also cut an immutable version
}));

ZipFile takes the raw bytes of the package; for large artifacts pass S3Bucket/S3Key instead.
Publish: true publishes a new numbered version from the updated code in one step.
For container-image functions, pass ImageUri rather than ZipFile.
As with config changes, the function goes Pending while the new code is deployed.

Related: Versions, Aliases & Traffic Shifting - what Publish: true gives you and how to route to it.

Intermediate Examples

5. Capture Tail Logs and Detect a Handler Error

Ask Lambda to return the last chunk of logs and check whether the function itself failed.

# --- Python (boto3) ---
import boto3, base64, json
 
lam = boto3.client("lambda")
resp = lam.invoke(
    FunctionName="my-func",
    InvocationType="RequestResponse",
    LogType="Tail",   # include last 4 KB of logs
    Payload=json.dumps({"name": "Ada"}).encode(),
)
if "FunctionError" in resp:
    print("handler failed:", resp["FunctionError"])
print(base64.b64decode(resp["LogResult"]).decode())

// --- TypeScript (AWS SDK v3) ---
import { LambdaClient, InvokeCommand } from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
const resp = await lam.send(new InvokeCommand({
  FunctionName: "my-func",
  InvocationType: "RequestResponse",
  LogType: "Tail",   // include last 4 KB of logs
  Payload: new TextEncoder().encode(JSON.stringify({ name: "Ada" })),
}));
if (resp.FunctionError) console.log("handler failed:", resp.FunctionError);
console.log(Buffer.from(resp.LogResult ?? "", "base64").toString());

LogType: "Tail" returns LogResult, base64-encoded, holding the last 4 KB of logs.
FunctionError is set (Unhandled or Handled) when the handler throws - even though the HTTP call itself succeeded.
A 200 status with a FunctionError is a successful invoke of a failing function - check both.
LogType only applies to synchronous invokes.

6. Wait Until an Update Finishes

Config and code updates are eventually consistent; wait for Active/Successful before the next step.

# --- Python (boto3) ---
import boto3
 
lam = boto3.client("lambda")
lam.update_function_configuration(FunctionName="my-func", MemorySize=1024)
 
waiter = lam.get_waiter("function_updated_v2")
waiter.wait(FunctionName="my-func")   # returns when LastUpdateStatus is Successful
print("update applied")

// --- TypeScript (AWS SDK v3) ---
import {
  LambdaClient,
  UpdateFunctionConfigurationCommand,
  waitUntilFunctionUpdatedV2,
} from "@aws-sdk/client-lambda";
 
const lam = new LambdaClient({});
await lam.send(new UpdateFunctionConfigurationCommand({ FunctionName: "my-func", MemorySize: 1024 }));
 
const r = await waitUntilFunctionUpdatedV2({ client: lam, maxWaitTime: 120 }, { FunctionName: "my-func" });
if (r.state !== "SUCCESS") throw new Error(`update not applied: ${r.state}`);
console.log("update applied");

The function_updated_v2 waiter polls GetFunctionConfiguration until LastUpdateStatus is Successful.
Chaining updates without waiting risks ResourceConflictException.
SDK v3's waitUntilFunctionUpdatedV2 returns a state object; branch on SUCCESS instead of catching.
There is a matching function_active_v2 / waitUntilFunctionActiveV2 for newly created functions.

FAQs

What status code should a synchronous invoke return?

200. An asynchronous (Event) invoke returns 202, and a DryRun returns 204. The status reflects the invoke itself, not whether your handler logic succeeded.

How do I tell if my handler threw an error?

Check for FunctionError in the response. If present (Unhandled or Handled), the handler failed even though the HTTP call returned 200. The payload then holds the error object.

Why is Payload bytes and not a string?

The SDKs return the raw response body. boto3 gives a streaming body you .read(); SDK v3 gives a Uint8Array you decode with TextDecoder. Then json.loads / JSON.parse it.

How do I get logs without CloudWatch?

Set LogType: "Tail" on a synchronous invoke to get the last 4 KB in LogResult, base64-encoded. For full history, read the function's CloudWatch Logs group.

Does update_function_configuration replace all settings?

Only the fields you pass, but structured fields like Environment.Variables are replaced wholesale. Merge in the keys you want to keep before sending.

Can I update code and publish a version at once?

Yes. Pass Publish: true to update_function_code, and Lambda publishes a new immutable version from the new code without a separate PublishVersion call.

Why do I get ResourceConflictException on back-to-back updates?

The previous change is still applying. Wait with the function_updated_v2 waiter until LastUpdateStatus is Successful, then issue the next update.

What is the maximum timeout and memory?

Timeout is up to 900 seconds (15 minutes). Memory ranges from 128 MB to 10,240 MB, and CPU scales with it.

Lambda's Execution Model for SDK Callers - the model behind these calls.
Versions, Aliases & Traffic Shifting - publishing and routing after you update code.
Layers & Shared Dependencies - attach shared code and dependencies via configuration.
Lambda via SDK Best Practices - the defaults to set for invokes and updates.

Stack versions: This page was written for boto3 1.43.x (Python 3.10+) and the AWS SDK for JavaScript v3 (Node.js 18+).

Lambda Deep Dive Basics

Buscar en la documentación

Lambda Deep Dive Basics