The Anthropic integration automatically instruments Anthropic SDK calls to capture performance data and errors for Claude models.
Installation
The integration is enabled by default in Node.js:
import * as Sentry from '@sentry/node';
Sentry.init({
dsn: 'your-dsn',
// anthropicAIIntegration is included by default
});
Basic Usage
Just use the Anthropic SDK normally:
import Anthropic from '@anthropic-ai/sdk';
const anthropic = new Anthropic({
apiKey: process.env.ANTHROPIC_API_KEY,
});
// Automatically instrumented
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
messages: [
{ role: 'user', content: 'What is the capital of France?' },
],
});
Configuration
Default Behavior
By default, prompts and responses are not captured:
Sentry.init({
dsn: 'your-dsn',
sendDefaultPii: false, // Default: no prompts/responses
});
Capture Prompts and Responses
Global Setting
Anthropic Only
Outputs Only
Enable for all AI integrations:Sentry.init({
dsn: 'your-dsn',
sendDefaultPii: true, // Captures all inputs/outputs
});
Enable only for Anthropic:Sentry.init({
dsn: 'your-dsn',
sendDefaultPii: false,
integrations: [
Sentry.anthropicAIIntegration({
recordInputs: true,
recordOutputs: true,
}),
],
});
Capture responses but not prompts:Sentry.init({
dsn: 'your-dsn',
integrations: [
Sentry.anthropicAIIntegration({
recordInputs: false,
recordOutputs: true,
}),
],
});
Integration Options
recordInputs
boolean
default:"sendDefaultPii"
Capture prompt messages
recordOutputs
boolean
default:"sendDefaultPii"
Capture completion responses
Captured Data
Always Captured
These attributes are always included:
{
'gen_ai.operation.name': 'chat',
'gen_ai.request.model': 'claude-3-5-sonnet-20241022',
'gen_ai.system': 'anthropic',
'gen_ai.usage.input_tokens': 20,
'gen_ai.usage.output_tokens': 95,
'gen_ai.response.finish_reasons': ['end_turn'],
'gen_ai.request.temperature': 1.0,
'gen_ai.request.max_tokens': 1024,
}
{
'gen_ai.system': 'You are a helpful AI assistant.',
'gen_ai.prompt.0.role': 'user',
'gen_ai.prompt.0.content': 'What is the capital of France?',
}
When recordOutputs: true
Completion responses are captured:
{
'gen_ai.completion.0.role': 'assistant',
'gen_ai.completion.0.content': 'The capital of France is Paris.',
'gen_ai.completion.0.finish_reason': 'end_turn',
}
Supported Operations
Message Creation
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
messages: [
{ role: 'user', content: 'Hello, Claude!' },
],
});
// Span: gen_ai.chat.completions
Streaming Messages
const stream = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
messages: [{ role: 'user', content: 'Tell me a story' }],
stream: true,
});
for await (const event of stream) {
if (event.type === 'content_block_delta') {
process.stdout.write(event.delta.text);
}
}
// Span includes full streaming response
System Prompts
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
system: 'You are a helpful customer service agent.',
messages: [
{ role: 'user', content: 'I need help with my order' },
],
});
// System prompt captured when recordInputs: true
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
tools: [
{
name: 'get_weather',
description: 'Get the current weather for a location',
input_schema: {
type: 'object',
properties: {
location: {
type: 'string',
description: 'City name',
},
},
required: ['location'],
},
},
],
messages: [
{ role: 'user', content: 'What is the weather in Paris?' },
],
});
// Tool use captured in span attributes
Practical Examples
Multi-Turn Conversation
import * as Sentry from '@sentry/node';
import Anthropic from '@anthropic-ai/sdk';
const anthropic = new Anthropic();
async function chat(conversationHistory) {
return await Sentry.startSpan(
{
name: 'Claude Chat',
op: 'ai.chat',
attributes: {
'conversation.turns': conversationHistory.length,
},
},
async () => {
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
messages: conversationHistory,
});
return message.content[0].text;
}
);
}
// Usage
const history = [
{ role: 'user', content: 'What is 2+2?' },
{ role: 'assistant', content: '2+2 equals 4.' },
{ role: 'user', content: 'What about 3+3?' },
];
const response = await chat(history);
Content Moderation
async function moderateContent(content) {
return await Sentry.startSpan(
{
name: 'Content Moderation',
op: 'ai.moderation',
},
async () => {
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 256,
system: 'You are a content moderation assistant. Analyze the following content and determine if it violates community guidelines.',
messages: [
{ role: 'user', content },
],
});
return {
safe: message.content[0].text.includes('safe'),
analysis: message.content[0].text,
};
}
);
}
Document Analysis
async function analyzeDocument(documentText) {
return await Sentry.startSpan(
{
name: 'Analyze Document',
op: 'ai.analysis',
attributes: {
'document.length': documentText.length,
},
},
async () => {
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 2048,
messages: [
{
role: 'user',
content: `Please analyze this document and provide a summary:\n\n${documentText}`,
},
],
});
return message.content[0].text;
}
);
}
async function getWeatherAssistant(userQuery) {
const tools = [
{
name: 'get_weather',
description: 'Get current weather for a location',
input_schema: {
type: 'object',
properties: {
location: { type: 'string' },
unit: { type: 'string', enum: ['celsius', 'fahrenheit'] },
},
required: ['location'],
},
},
];
return await Sentry.startSpan(
{ name: 'Weather Assistant', op: 'ai.tool_use' },
async () => {
const message = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
tools,
messages: [{ role: 'user', content: userQuery }],
});
// Check if Claude wants to use a tool
const toolUse = message.content.find(block => block.type === 'tool_use');
if (toolUse) {
// Call the actual weather API
const weatherData = await fetchWeather(toolUse.input.location);
// Continue conversation with tool result
const followUp = await anthropic.messages.create({
model: 'claude-3-5-sonnet-20241022',
max_tokens: 1024,
tools,
messages: [
{ role: 'user', content: userQuery },
{ role: 'assistant', content: message.content },
{
role: 'user',
content: [
{
type: 'tool_result',
tool_use_id: toolUse.id,
content: JSON.stringify(weatherData),
},
],
},
],
});
return followUp.content[0].text;
}
return message.content[0].text;
}
);
}
Model Support
The integration supports all Claude models:
- Claude 3.5 Sonnet:
claude-3-5-sonnet-20241022
- Claude 3 Opus:
claude-3-opus-20240229
- Claude 3 Sonnet:
claude-3-sonnet-20240229
- Claude 3 Haiku:
claude-3-haiku-20240307
- Claude 2.1:
claude-2.1
- Claude 2.0:
claude-2.0
Track Claude performance metrics:
- Response Times: API latency per model
- Token Usage: Input and output tokens
- Stop Reasons: Track why generations ended
- Error Rates: Rate limits and API errors
Source Code
The Anthropic integration is implemented in:
packages/node/src/integrations/tracing/anthropic-ai/index.ts:11
Privacy Best Practices
Claude conversations may contain sensitive user information.
Sentry.init({
dsn: 'your-dsn',
sendDefaultPii: true,
beforeSendSpan(span) {
// Remove PII from prompts
const promptContent = span.attributes?.['gen_ai.prompt.0.content'];
if (promptContent && typeof promptContent === 'string') {
// Redact email addresses
span.attributes['gen_ai.prompt.0.content'] = promptContent
.replace(/[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}/g, '[EMAIL]');
}
return span;
},
});
Selective Capture by Use Case
async function processQuery(query, options = {}) {
const isPublic = options.isPublic ?? false;
// Use a custom client configuration based on sensitivity
const integrations = isPublic
? [Sentry.anthropicAIIntegration({ recordInputs: true, recordOutputs: true })]
: [Sentry.anthropicAIIntegration({ recordInputs: false, recordOutputs: false })];
// Process with appropriate privacy settings
const message = await anthropic.messages.create({...});
return message;
}
Troubleshooting
Spans Not Appearing
Ensure tracing is enabled:
Sentry.init({
dsn: 'your-dsn',
tracesSampleRate: 1.0, // Capture 100% of traces
});
Streaming Responses Not Captured
Streaming responses are fully captured. Token usage appears after the stream completes.
Tool use is captured in span attributes. Enable recordOutputs: true to see tool calls:
Sentry.anthropicAIIntegration({
recordOutputs: true, // Captures tool use blocks
});
