AWS Lambda Integration
Observe serverless function invocations, durations, cold starts, and errors in real time. Correlate Lambda metrics with distributed traces and get AI-powered anomaly detection for every function.
How It Works
Deploy CloudFormation Stack
Launch the TigerOps Lambda monitoring stack. It creates the IAM role, CloudWatch Metric Streams for the AWS/Lambda namespace, and a Lambda Extension layer for enhanced metrics.
Add Lambda Extension Layer
Attach the TigerOps Lambda Extension layer ARN to your functions. The extension collects enhanced runtime metrics — including init duration, memory usage, and cold start frequency — without modifying your function code.
Enable X-Ray or OTEL Tracing
Enable AWS X-Ray active tracing or the OpenTelemetry Lambda layer. TigerOps ingests traces and correlates them with the invocation metrics so you can see latency breakdowns per function invocation.
Set Function-Level Alerts
Configure per-function error rate, throttle rate, and p99 duration alerts in TigerOps. Alerts fire before timeout budgets are exhausted and include the correlated trace for the slowest invocations.
What You Get Out of the Box
Cold Start Tracking
Track cold start frequency and init duration per function and runtime. TigerOps identifies functions with excessive cold starts and correlates them with deployment or concurrency changes.
Invocation & Error Rates
Invocation count, success rate, error rate, and throttle rate per function and alias. Trend charts and anomaly detection surface sudden error spikes before they escalate.
Duration Percentiles
p50, p95, and p99 duration tracking per function. TigerOps alerts when p99 duration approaches the function timeout and predicts timeout breaches using trend analysis.
Concurrency & Throttle Visibility
Reserved and provisioned concurrency utilisation alongside throttle rates. TigerOps warns when concurrency approaches account limits and suggests provisioned concurrency thresholds.
Memory & Cost Optimisation
Memory allocation vs. actual peak usage per function. TigerOps surfaces over-provisioned functions and estimates the monthly cost saving from right-sizing memory configurations.
Distributed Trace Correlation
Every Lambda invocation metric is linked to the corresponding distributed trace. Jump from an error spike to the exact failed invocation trace with full upstream and downstream context.
Lambda Extension Layer Setup
Attach the TigerOps extension layer to your Lambda functions for enhanced metrics.
# Attach TigerOps Lambda Extension layer
LAYER_ARN="arn:aws:lambda:us-east-1:123456789:layer:tigerops-extension:latest"
aws lambda update-function-configuration \
--function-name my-function \
--layers "${LAYER_ARN}" \
--environment "Variables={TIGEROPS_API_KEY=${TIGEROPS_API_KEY},TIGEROPS_INGEST=https://ingest.atatus.net}"
# Or via CloudFormation / SAM template
# template.yaml
Resources:
MyFunction:
Type: AWS::Serverless::Function
Properties:
FunctionName: my-function
Handler: index.handler
Runtime: nodejs20.x
Layers:
- !Sub "arn:aws:lambda:${AWS::Region}:123456789:layer:tigerops-extension:latest"
Environment:
Variables:
TIGEROPS_API_KEY: !Ref TigerOpsApiKey
TIGEROPS_INGEST: https://ingest.atatus.net
Tracing: Active # Enable X-Ray for trace correlation
# Enable CloudWatch Metric Streams for AWS/Lambda namespace
aws cloudformation deploy \
--template-url https://tigerops-cfn.s3.amazonaws.com/lambda-integration.yaml \
--stack-name tigerops-lambda \
--capabilities CAPABILITY_IAM \
--parameter-overrides TigerOpsApiKey=${TIGEROPS_API_KEY}Common Questions
What metrics does the TigerOps Lambda Extension provide beyond CloudWatch?
The extension provides init duration breakdown (import time vs. handler init), memory peak usage per invocation, extension overhead time, and per-invocation cold/warm status — none of which are available in standard CloudWatch Lambda metrics.
Does adding the Lambda Extension layer increase my function latency?
The TigerOps Lambda Extension runs as a separate process in the Lambda execution environment and operates asynchronously during the IDLE phase. It adds zero latency to your function handler and less than 5 ms overhead on shutdown.
Can TigerOps monitor Lambda functions across all AWS regions?
Yes. Deploy the TigerOps CloudFormation stack in each region and all function metrics are aggregated into a single TigerOps workspace. You can filter by region, account, or function name from the unified dashboard.
How does TigerOps handle Lambda function versions and aliases?
TigerOps tracks metrics per function alias ($LATEST, production, canary) and version. This allows you to compare performance between a canary and the production alias during a weighted alias deployment.
Can TigerOps alert on Lambda timeout errors specifically?
Yes. TigerOps distinguishes between timeout errors, out-of-memory errors, and code-level errors from the CloudWatch Logs error patterns and Lambda extension data. Each error type has a separate alert condition you can configure independently.
See Every Lambda Invocation, Cold Start, and Error
Real-time serverless observability with trace correlation and AI anomaly detection. Zero code changes required.