canonical: https://jentic.com/apis/amazonaws.com/aws-application-insights

# AWS Amazon CloudWatch Application Insights

CloudWatch Application Insights detects issues across the components of an application by analysing CloudWatch metrics, logs, and events. It builds an application model from resource groups, recognises common workloads such as SQL Server and .NET, and surfaces likely root causes as observations and problems.

## For AI agents

Use the Amazon CloudWatch Application Insights API to detect application issues across CloudWatch metrics, logs, and events on AWS, with 27 operations covering the full control-plane lifecycle.

## Scope

Does not handle metric ingestion, log storage, or distributed tracing - use for cross-component problem detection on top of existing CloudWatch data only.

## Capabilities

- Onboard a resource group as an Application Insights application with auto-configured component detection
- Define components that group EC2 instances, RDS databases, or other resources by role
- Configure log patterns that classify CloudWatch Logs entries by severity
- Browse detected problems with correlated observations across components
- Inspect observations that link metrics, logs, and CloudWatch events to a problem
- Update component configuration to refine monitoring rules over time

## Use cases

### Automated problem detection for a Windows .NET stack

Onboard a resource group containing IIS web servers, .NET application servers, and a SQL Server backend to Application Insights, which then auto-detects the workload and configures CloudWatch metrics, logs, and events monitoring across the components. Detected problems combine signals from all components into a single root-cause narrative. Initial onboarding is typically under an hour.

Example prompt: Create an Application Insights application against the 'orders-prod' resource group, then list any detected problems with their observation counts.

### Log-pattern-driven incident classification

Create custom log patterns that match application-specific error strings - for example custom checkout exceptions or database deadlock messages - and assign severity levels. Application Insights then incorporates pattern matches into its problem detection and observation feeds, raising the priority of incidents that match high-severity patterns. Configuration takes minutes per pattern.

Example prompt: Create a log pattern named 'checkout-deadlock' with the regex 'CheckoutDeadlockException' and severity HIGH attached to the orders-prod application's pattern set.

### Post-deployment regression check

After a new deployment, list the application's detected problems, filter for ones with start times after the deployment, and review their contributing observations to identify whether the release introduced regressions. This automates the manual triage step of comparing dashboards before and after a release. Useful for any team running CloudWatch-monitored workloads.

Example prompt: List all problems for the orders-prod application started in the last hour, then describe observations for any with status 'Open'.

### AI agent triaging Application Insights problems through Jentic

An on-call agent can investigate detected problems autonomously by calling Application Insights through Jentic. It searches for problem-listing intents, loads the schemas for ListProblems and DescribeProblemObservations, and executes them with scoped credentials. The agent surfaces a ranked summary of active problems without engineers having to sign in to the CloudWatch console.

Example prompt: Search Jentic for 'list detected problems for an application', load the schemas for ListProblems and DescribeProblemObservations, and execute them against the orders-prod application.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.CreateApplication | Create an Application Insights application |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.CreateComponent | Add a component to an application |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.CreateLogPattern | Create a log pattern for log classification |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.DescribeProblem | Describe a detected problem |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.ListApplications | List Application Insights applications |
| POST | /#X-Amz-Target=EC2WindowsBarleyService.DescribeProblemObservations | List observations contributing to a problem |

## Key resources

- **Applications** — Application models built on top of a resource group with associated components and configuration
- **Components** — Logical groupings of resources such as a SQL Server instance, IIS server, or load balancer
- **LogPatterns** — Regex patterns used to classify log events by severity
- **Problems** — Detected issues correlating metrics, logs, and events across components
- **Observations** — Individual signals that contribute to a detected problem

## Why Jentic

- **Setup:** Wiring Amazon CloudWatch Application Insights by hand means creating IAM credentials, choosing the right regional host from application-insights.{region}.amazonaws.com, signing every request with AWS SigV4, and passing the X-Amz-Target action header on each call yourself. Through Jentic you install once, import Amazon CloudWatch Application Insights from the API Directory, store the AWS access key once, and your agent calls it.
- **Permission scoping:** Amazon CloudWatch Application Insights carries the application and component targets in the request body rather than the URL path, so limit the agent to the operations it needs, such as CreateApplication, DescribeProblem, or ListApplications. You choose the operations it may call, so destructive ones like deleting an application or removing a component are not included unless you add them.
- **Credential handling:** Your AWS access key for Amazon CloudWatch Application Insights is stored once, encrypted, by your own Jentic One instance and signed with SigV4 at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'detect problems across my application's components' or 'describe observations for a detected problem', and Jentic returns the matching Amazon CloudWatch Application Insights operation with its input schema so the agent calls the right endpoint without browsing the AWS service reference.

## Related APIs

- **CloudWatch Logs** — CloudWatch Logs supplies the log data that Application Insights classifies and correlates
- **Distributed tracing** — AWS X-Ray traces requests across distributed services rather than detecting problems on existing CloudWatch data
- **Security findings aggregation** — Security Hub aggregates security findings while Application Insights aggregates operational problems

## FAQ

### What authentication does the Amazon CloudWatch Application Insights API use?

The Amazon CloudWatch Application Insights API uses AWS Signature Version 4 (HMAC) request signing with IAM-issued credentials, the same scheme as every AWS service API. Jentic's your Jentic One instance stores those AWS credentials encrypted in the vault, generates short-lived signed requests at execution time, and never passes raw access keys into the agent's context.

### Can I onboard my resource group to application insights with the Amazon CloudWatch Application Insights API?

Yes - the Amazon CloudWatch Application Insights API exposes 27 operations including the actions needed for that scenario. Use the operations listed in the key endpoints section as the starting point, then chain calls as needed for your workflow.

### What are the rate limits for the Amazon CloudWatch Application Insights API?

AWS applies per-account, per-region request rate limits to the Amazon CloudWatch Application Insights control plane. Specific limits are not encoded in the OpenAPI spec; consult the AWS service quotas console for the Amazon CloudWatch Application Insights entry, and design retries with exponential backoff to absorb throttling responses.

### How do I detect problems across my application's components through Jentic?

Run pip install jentic, then call client.search('detect problems across my application's components') to discover the Amazon CloudWatch Application Insights operations that match. Load the schema for the chosen operation with client.load(...) and execute it with client.execute(...). Jentic handles AWS request signing automatically against the credentials stored in your Jentic One instance.

### Is the Amazon CloudWatch Application Insights API free to call?

AWS does not charge for control-plane API calls themselves on most Amazon CloudWatch Application Insights operations, but the underlying resources you create or operate (fleets, queries, deployments, and so on) incur usage charges according to the Amazon CloudWatch Application Insights pricing page. Refer to the AWS pricing page for the service to estimate cost.

### Which operations should an agent call first when working with the Amazon CloudWatch Application Insights API?

For most workflows, agents should start by listing existing resources to understand the current state, then call the create or update operation that matches the intent. The endpoints listed under Key Endpoints in the catalog give a ranked starting set.

### Can I limit what my agent is allowed to do with the AWS CloudWatch Application Insights API?

Yes. Because you run Jentic One yourself, your own rules decide which of the API's operations and which AWS credentials the agent may use. Since Application Insights carries the application and component targets in the request body rather than the URL path, you scope by operation: allow read and setup calls like ListApplications, DescribeProblem, and CreateApplication while withholding destructive ones such as deleting an application or removing a component. The agent can only invoke the operations you have explicitly granted, so anything you do not add stays off-limits.