Amazon CodeGuru Profiler
aws.amazon.com
- Free plan
- Linux
- Web
- API
- Self-hosted
What Amazon CodeGuru Profiler is
Amazon CodeGuru Profiler collects runtime performance data from live applications to help teams find CPU bottlenecks and improve performance. It is designed for continuous production profiling with minimal overhead and presents CPU and latency analysis in interactive flame graphs, showing which code paths use the most resources. Machine learning recommendations identify expensive methods and describe their cause, impact and possible resolution. The service can detect anomalies in application profiles and send an Amazon SNS notification when configured. It supports Java and JVM languages such as Scala and Kotlin, plus Python 3.6 or later. Java and other JVM languages can be profiled for CPU and heap summary; Python profiling covers CPU wall-clock time. Applications may run on EC2, ECS, EKS, Fargate, Lambda or on premises. An AWS account is required, with IAM permissions controlling access. The agent collects stack traces, not function parameters, variable values or application data. Profiles are retained for periods ranging from 15 days for five-minute profiles to three years for daily profiles. The Free Tier lasts 90 days; afterward, standard sampling-hour pricing applies.
Who it suits
It suits teams investigating runtime performance in supported Java, JVM or Python applications. Users need an AWS account, and the profiling detail available differs by language.
Good to know, and what to watch
- Continuous production profiling with minimal overhead
- Interactive CPU and latency flame graphs
- Recommendations explain cause, impact and resolution
- Anomaly detection can notify through Amazon SNS
- Stack traces collected without application data
- Requires an AWS account
- Python profiling covers CPU wall-clock time
- Free Tier lasts 90 days
- Standard sampling-hour pricing applies afterward
Plans
Every plan and priceAmazon CodeGuru Profiler Free Tier
Amazon CodeGuru Profiler on AWS Lambda
Amazon CodeGuru Profiler on EC2, ECS, EKS, and AWS Fargate
90-day trial
Getting started
- Visit https://aws.amazon.com/codeguru/profiler/.
- Use or create an AWS account.
- Configure IAM permissions for people who need to create or access profiler resources.
- Enable the profiling agent in a JVM application from the command line or application code, or use it to profile a Lambda application.
- Configure Amazon SNS notifications if you want alerts about new anomalies.
Limits to check
The free tier lasts 90 days; standard sampling-hour pricing applies afterward. For Java and other JVM languages, profiling covers CPU and heap summary; Python profiling covers CPU wall-clock time. The EC2, ECS, EKS, and Fargate plan charges $0.005 per sampling hour for the first 36,000 sampling hours per profiling group per month, with no additional charge beyond that. The Lambda plan gives the first 500 sampling hours per payer account free, then charges $0.005 per sampling hour up to 36,000 hours per profiling group per month, with no additional charge beyond that.
The Club's review
Amazon CodeGuru Profiler is a cloud service for monitoring runtime performance in live applications. It suits teams running supported Java, JVM, or Python workloads that need ongoing CPU and latency insight. Its production profiling and actionable profile views make it a practical AWS option, though Python’s narrower profiling scope and sampling-hour charges matter.
Overview
CodeGuru Profiler samples running applications continuously, with minimal overhead, to help teams find CPU bottlenecks and improve performance. That focus on production behavior is useful when issues emerge under real workloads. Teams that need a profiling mode other than sampling should look elsewhere.
The service requires an AWS account, and IAM permissions govern who can create or access its resources. The agent collects stack traces but does not access function parameter names or values, variable values, or application data. Data is encrypted at rest and in transit, and TLS protects traffic to the service and downstream dependencies. AWS says third-party auditors assess CodeGuru Profiler under compliance programs including SOC, PCI, FedRAMP, and HIPAA.
Key features
Interactive flame graphs display CPU and latency analysis by code path, helping developers locate work that consumes resources. Machine-learning recommendations identify costly methods and explain their cause, impact, and possible resolution. Those explanations can help teams prioritize investigation, while the team still needs to decide whether a proposed change suits its application.
Anomaly detection flags changes in application profiles and can send an Amazon SNS notification when a new anomaly appears, if configured. This gives teams a way to surface profile changes without continuously watching the graphs.
Profiling uses sampling and covers CPU, memory, and threads. Java and other JVM languages support CPU and heap summaries; Python 3.6 or later is profiled for CPU wall-clock time. The profiling agent can be enabled in JVM applications from the command line or application code and is also used for Lambda applications. This makes the service most useful where its language-specific coverage matches the question a team needs to answer.
Profiles are retained at different resolutions: five-minute profiles for 15 days, hourly profiles for 60 days, and daily profiles for three years. That long daily history can help with longer-term comparisons, while the shorter windows limit how far back teams can inspect finer-grained profiles.
Pricing
The Amazon CodeGuru Profiler Free Tier costs nothing for 90 days. After that, standard sampling-hour pricing applies, so teams should account for ongoing usage once the free period ends.
For EC2, ECS, EKS, and AWS Fargate, the plan is priced at $0.01 a month, billed per sampling hour. The first 36,000 sampling hours per profiling group per month cost $0.005 per sampling hour; there is no additional charge beyond 36,000 sampling hours per profiling group per month. This structure suits teams profiling supported workloads on those services, though usage within the priced band drives the bill.
For AWS Lambda, the plan is priced at $0.01 a month. The first 500 sampling hours per payer account are free, then sampling costs $0.005 per hour up to 36,000 sampling hours per profiling group per month, with no additional charge beyond that. The account-level free allowance gives Lambda users a small initial buffer; higher use incurs sampling charges.
Platforms
CodeGuru Profiler is a cloud service with API, Linux, self-hosted, and web platform coverage. Supported applications can run on EC2, ECS, EKS, Fargate, Lambda, or on premises. Supported languages are Java and other JVM languages, plus Python 3.6 or later, with the profiling scope differing by language.
Who it's for
CodeGuru Profiler is a strong fit for development and operations teams that run supported workloads and want continuous production profiles, flame graphs, anomaly alerts, and recommendations in one service. Its retention windows also suit teams that need to look back at daily profile history. It is a weaker fit for applications outside its supported languages or for teams that need equivalent CPU and heap profiling across Java and Python.
Pros and cons
- Pros: Continuous production profiling with minimal overhead helps teams observe live behavior.
- Pros: Flame graphs, method recommendations, and configurable SNS anomaly notifications connect profile data to investigation and alerting.
- Pros: Daily profiles remain available for three years, supporting longer-term review.
- Pros: The agent collects stack traces without accessing parameter or variable values or application data.
- Cons: Python profiling covers CPU wall-clock time, while Java and JVM languages also receive heap summaries.
- Cons: Sampling-hour charges apply after the 90-day Free Tier, with pricing tied to the workload and usage band.
- Cons: An AWS account and IAM permissions are required, which makes it less suitable for teams that do not use AWS access controls.
Alternatives
Profiling Software and Continuous Profiling Tools are useful starting points for comparing products in those categories.
Grafana Cloud Application Observability is worth considering for teams seeking a freemium option with an always-free allowance of 2,232 host hours and a Pro plan billed per host hour.
YourKit .NET Profiler is an alternative for readers evaluating a paid profiler with a free license option for open-source project contributors and Microsoft MVPs.
YourKit Java Profiler is another paid Java profiling option, with a single-seat annual subscription listed at $449,449 a year for one installed copy and one user at a time.
Intel VTune Profiler is a free download for readers looking for a standalone tool or one available through the Intel oneAPI Toolkit.
Javadoc is included with JDK 25, making it a no-cost option for readers whose needs match that tool.
Node.js --prof is a free option for generating V8 profiler output.
Parca is a free, open-source profiling option with components available under the Apache 2 License.
AMD Auto-Detect and Install Tool is a free Windows tool for supported AMD Radeon graphics or Ryzen systems.
Verdict
Choose Amazon CodeGuru Profiler if your team runs supported Java, JVM, or Python applications and wants continuous production sampling, recommendations, and profile anomaly alerts within an AWS account. Its clearest advantage is the combination of long-lived profile history and tools that help turn sampled data into investigation. Look elsewhere if Python heap profiling, a non-sampling method, or avoiding AWS and usage-based charges is central to your choice.
The verdict
CodeGuru Profiler brings production profiling, flame graphs, recommendations and anomaly detection together for supported applications. Teams should check the language-specific profiling scope and account for standard sampling-hour pricing after the 90-day Free Tier.
Questions people ask
Is there a free plan or trial?
The Amazon CodeGuru Profiler Free Tier costs nothing for 90 days. After that, standard sampling-hour pricing applies.
What does it cost after the free tier?
For EC2, ECS, EKS, and Fargate, the listed rate is $0.005 per sampling hour for the first 36,000 sampling hours per profiling group per month; there is no additional charge beyond 36,000 hours. For Lambda, the first 500 sampling hours per payer account are free, then the same $0.005 rate applies up to 36,000 hours per profiling group per month.
Which languages does it support?
It supports Java, JVM languages such as Scala and Kotlin, and Python 3.6 or later. Java and JVM languages provide CPU and heap summary profiling; Python profiling covers CPU wall-clock time.
Where can profiled applications run?
The FAQ lists EC2, ECS, EKS, Fargate, Lambda, and on-premises environments.
Does it send alerts about anomalies?
It can send an Amazon SNS notification when it detects a new anomaly, if notifications are configured.
What data does the agent collect?
The agent collects stack traces. It does not access function parameter names or values, variable values, or application data.
What the maker publishes
- Anomaly detection
- It detects anomalies in application profiles and can send an Amazon SNS notification when a new anomaly is detected if configured. source
- Compliance
- AWS says third-party auditors assess CodeGuru Profiler as part of multiple compliance programs, including SOC, PCI, FedRAMP, and HIPAA. source
- Data collected
- The agent collects stack traces but does not access function parameter names or values, variable values, or application data. source
- Data retention
- The service retains five-minute profiles for 15 days, hourly profiles for 60 days, and daily profiles for three years. source
- Encryption
- Data at rest and in transit are encrypted by default; traffic to the service and downstream dependencies uses TLS. source
- Flame graphs
- It presents CPU and latency analysis in interactive flame graphs to show which code paths consume the most resources. source
- Integration
- The profiling agent can be enabled in JVM applications from the command line or application code and is also used to profile Lambda applications. source
- Language capability limit
- The FAQ says Java and other JVM languages can be profiled for CPU and heap summary, while Python profiling covers CPU wall-clock time. source
- Production profiling
- The profiler is designed to run continuously on production applications with minimal overhead. source
- Purpose
- Amazon CodeGuru Profiler collects runtime performance data from live applications and provides recommendations to improve performance and find CPU bottlenecks. source
- Recommendations
- Machine learning recommendations identify expensive or resource-intensive methods and explain the cause, impact, and possible resolution. source
- Runtime support
- The FAQ lists Java, Python, and JVM languages such as Scala and Kotlin, and says applications can run on EC2, ECS, EKS, Fargate, Lambda, or on premises. source
- Support and permissions
- Using the service requires an AWS account, and IAM permissions control who can create or access CodeGuru Profiler resources. source
Checked against the maker's pages on 8 Oct 2026.