Amazon CloudWatch
Cloud service · AWS · Updated

Amazon CloudWatch

In short

Amazon CloudWatch is a monitoring and observability service offered by Amazon Web Services (AWS) that provides data and actionable insights to monitor applications, respond to system-wide performance changes, optimize resource utilization, and get a unified view of operational health.

Cloud Services Hub →

8 key features of Amazon CloudWatch

Amazon CloudWatch provides real-time monitoring, customizable alarms, dashboards for consolidated data visualization, logs management, event-driven automation, the capability to publish custom metrics, and seamless integration with AWS services.

01
Real-time Monitoring

Amazon CloudWatch enables real-time monitoring of AWS resources and applications, allowing users to collect and track metrics, collect and monitor log files, and set alarms.

02
Alarms

Users can set alarms in CloudWatch to receive notifications or automatically change the resources being monitored, based on predefined rules, which helps in proactive issue resolution and automation of actions.

03
Dashboards

CloudWatch offers customizable dashboards that provide a consolidated view of the metrics and alarms for your AWS resources, making it easier to view and analyze your operations.

04
Logs

With CloudWatch Logs, users can monitor, store, and access their log files from Amazon EC2 instances, AWS CloudTrail, and other sources, enabling efficient troubleshooting and security analysis.

05
Amazon EventBridge

Amazon EventBridge (formerly Amazon CloudWatch Events) deliver a near real-time stream of system events that describe changes in AWS resources, allowing users to respond automatically to state changes with triggered actions.

06
Custom Metrics

Users can publish their own metrics to CloudWatch, enabling them to monitor application performance and operational health uniquely suited to their needs.

07
Integration

CloudWatch seamlessly integrates with other AWS services, providing comprehensive monitoring solutions that enhance the performance and availability of cloud resources.

08
Automatic Service Dashboards

The CloudWatch home page automatically displays metrics about every AWS service you use.

Where Amazon CloudWatch is used

Amazon CloudWatch is utilized for application and infrastructure monitoring, log management and analysis, automated alarming and notifications, and creating custom dashboards for comprehensive operational visibility.

  1. 01
    Application Monitoring

    Amazon CloudWatch facilitates real-time monitoring of applications by gathering and tracking metrics, which enables teams to observe application performance, optimize resource utilization, and identify and respond to issues promptly, ensuring high availability and performance.

  2. 02
    Infrastructure Monitoring

    This use case involves using CloudWatch to monitor the health and performance of AWS resources such as EC2 instances, DynamoDB tables, and RDS database instances. It helps in detecting failures, measuring resource utilization, and ensuring that the infrastructure scales properly to meet demand.

  3. 03
    Log Management and Analysis

    CloudWatch Logs support the collection, monitoring, and analysis of log files from AWS resources, applications, and on-premises servers. This facilitates troubleshooting, the security monitoring of environment, and the retention of logs for compliance purposes.

  4. 04
    Automated Alarming and Notifications

    With CloudWatch, users can set up alarms and notifications based on specific metrics or logs. This helps in automated incident detection and enables teams to take immediate action or automate responses to ensure continuous operation and mitigate issues.

  5. 05
    Custom Dashboards for Operational Visibility

    CloudWatch allows the creation of customizable dashboards to visualize the performance and health of AWS resources and applications. This centralized view helps stakeholders understand operational status and make informed decisions.

How Amazon CloudWatch is priced

Amazon CloudWatch pricing includes a free tier for basic monitoring, a pay-as-you-go model for additional usage, and extra charges for detailed monitoring.

Amazon CloudWatch Free Tier

Offers a limited amount of data ingestion, alarm evaluations, and dashboard metrics for free each month.

Pay-as-you-go Pricing

There is no upfront commitment or minimum fee; charges are based on usage and billed at the end of the month.

Detailed Monitoring

Additional charges apply for enabling detailed monitoring for Amazon EC2 instances and other AWS services, providing more granular data.

Classic Custom Metrics

Custom metric charges are prorated hourly and incurred only during hours when metrics are sent; each unique combination of dimensions is treated as a separate metric.

CloudWatch Omni

Standard CloudWatch ingestion and storage rates apply, with additional charges for queries above the included allowance.

OpenTelemetry Metrics

Publishing OpenTelemetry metrics costs $0.50 per GB ingested, including 15 months of storage with no separate API-call, metric-storage, or unique-series charges.

OpenTelemetry Vended Metric Enrichment

Converting AWS-generated CloudWatch metrics to OpenTelemetry format and enriching them with resource tags and account metadata costs $0.50 per GB ingested.

PromQL Queries

PromQL API queries cost $0.01 per million samples scanned; queries through the CloudWatch console, including dashboards, are free, while PromQL alarm evaluations incur sample-scanning query charges.

Logs Insights Queries

Logs Insights queries are charged according to the amount of ingested log data scanned by each query.

Logs Storage

Log storage is billed per compressed GB, with separate rates for Standard, Infrequent Access, and Archive Instant Access storage.

Logs Live Tail

Live Tail costs $0.01 per minute.

Logs Data Protection

Data Protection is free for Standard logs and charged per GB scanned for Infrequent Access logs.

Log Centralization

For cross-account and cross-Region log centralization, data transfer and ingestion for the first copy are free; additional copies cost $0.05 per GB, and storage charges apply to every copy.

Data Transfer

CloudWatch has no data transfer IN charge; data transfer OUT from CloudWatch Logs follows the corresponding Amazon EC2 outbound transfer rates.

Vended Logs

Logs from designated AWS services use volume-tiered pricing and can be delivered to CloudWatch Logs, Amazon S3, or Amazon Data Firehose.

Analyze with OpenSearch

Analyzing CloudWatch Logs through OpenSearch incurs standard CloudWatch Logs ingestion and storage charges plus OpenSearch Serverless and OpenSearch Direct Query usage charges.

Contributor Insights for CloudWatch Logs

Contributor Insights log rules are charged per rule per month and per million matching log events per month.

Contributor Insights for DynamoDB

Contributor Insights for DynamoDB is charged per rule per month and per million DynamoDB events generated when table items are read or written.

Contributor Insights for PrivateLink

Contributor Insights for PrivateLink is charged per rule per month, with each rule evaluating up to 50,000 endpoints.

Metric Streams

Metric Streams are billed by metric update, including four default statistics; each additional set of five statistics requested for a metric counts as another update, and Amazon Data Firehose ingestion charges also apply.

Metric Centralization

The first physical metric copy to a destination account is free; subsequent copies to additional Regions or accounts cost $0.05 per GB copied.

Dashboards

Custom dashboard charges are prorated hourly for as long as each dashboard exists.

Alarms

Alarm charges accrue hourly from the first hour an alarm exists; metric alarms are charged for each metric listed in the alarm expression.

Application Signals and Transaction Search

The bundled Application Signals and transaction search model charges per GB of application metrics and spans ingested and per million spans indexed as X-Ray trace summaries, with the first 1% of spans indexed free.

Application Signals Without Transaction Spans

Application Signals without transaction-span visibility is billed per million signals: each inbound or outbound application request generates one signal, and each SLO generates two signals per SLI metric period.

X-Ray Traces

X-Ray charges separately for traces recorded, retrieved, and scanned.

Synthetics

Synthetics is billed per canary run.

Real User Monitoring

Web RUM is billed by event count, while RUM OpenTelemetry spans and events are billed per GB ingested.

Container Insights with OpenTelemetry for EKS

Container Insights with OpenTelemetry for EKS costs $0.08 per GB ingested, with no separate API-call, storage, or metric-series charges.

Container Insights Enhanced Observability for EKS

Enhanced observability for EKS uses tiered observation-based pricing that combines metric storage and performance-log ingestion; log storage and container-log ingestion remain separately charged.

Container Insights Enhanced Observability for ECS

Enhanced observability for ECS is billed per metric per month, prorated hourly, combining metric storage and performance-log ingestion; log storage and container-log ingestion remain separately charged.

Database Insights Retention

Database Insights retention charges use vCPU-hours for provisioned instances and ACU-hours for Serverless and Limitless instances, with separate first-month and additional-month retention rates.

Custom Logs Ingestion

Custom log ingestion is billed per GB, with separate rates for Standard and Infrequent Access.

Classic Metric API Usage

PutMetricData incurs request charges in addition to metric storage; GetMetricData and GetInsightRuleReport are billed by metrics requested, while GetMetricStatistics and ListMetrics are billed by requests.

Metrics Insights Queries

Metrics Insights queries through GetMetricData are billed by the total number of metrics analyzed; Metrics Insights queries through the CloudWatch console are free.

Log Query Alarms

Log query alarms incur a standard alarm charge plus query charges per GB scanned.

Metrics Insights Query Alarms

Metrics Insights query alarms are billed for each metric analyzed by the query.

Anomaly Detection Alarms

Anomaly detection alarms are billed for each metric in the alarm expression plus two additional alarm metrics for the model's upper and lower bands.

Composite Alarms

Composite alarms are billed per composite alarm per month.

Dashboard API Usage

GetDashboard, ListDashboards, PutDashboard, and DeleteDashboards are billed per API request.

Services Amazon CloudWatch works with

Amazon EC2Amazon EC2

Monitors the performance and resource utilization of Amazon EC2 instances, including CPU utilization, disk read/write operations, and network traffic.

Amazon RDSAmazon RDS

Tracks the operational health and performance metrics for Amazon RDS instances, including CPU utilization, storage space, and memory usage.

Amazon S3Amazon S3

Provides storage metrics and counts HTTP requests for Amazon S3 buckets, including the number of bytes downloaded or uploaded.

AWS LambdaAWS Lambda

Monitors AWS Lambda functions by tracking metrics such as invocation count, duration, and error rates.

Amazon API GatewayAmazon API Gateway

CloudWatch collects API Gateway request counts, latency, and error metrics, along with configured access and execution logs. Alarms can flag rising API error rates or latency.

Amazon AppFlowAmazon AppFlow

CloudWatch receives AppFlow flow-execution metrics, including records processed and execution failures, so you can monitor data transfers and alarm on unsuccessful runs.

Amazon AthenaAmazon Athena

CloudWatch receives Athena query metrics for enabled workgroups, including execution time and data scanned. Athena can also query CloudWatch telemetry exposed through the S3 Tables integration.

Amazon Augmented AI (Amazon A2I)Amazon Augmented AI (Amazon A2I)

CloudWatch receives Amazon A2I human-review workflow metrics, allowing you to track human-loop activity and monitor review completion times.

Amazon AuroraAmazon Aurora

CloudWatch collects Aurora database and cluster metrics, including connection counts, replica lag, and storage activity. Supported database logs and Enhanced Monitoring output can also be delivered to CloudWatch Logs.

Amazon Aurora ServerlessAmazon Aurora Serverless

CloudWatch tracks Aurora Serverless capacity and utilization metrics, including serverless database capacity and ACU utilization. These metrics help identify workloads approaching their configured capacity limits.

What is Amazon CloudWatch

4 min read

Amazon CloudWatch is a versatile monitoring and observability service designed by AWS (Amazon Web Services) to help IT professionals, developers, site reliability engineers (SREs), and DevOps teams track and manage the health, performance, and overall activity of their AWS resources and applications, both on AWS infrastructure and on-premises. Serving as a central component of AWS's extensive ecosystem, CloudWatch provides a detailed view into operational health in real-time by collecting monitoring and operational data in the form of logs, metrics, and events, making it an indispensable tool for maintaining the performance and smooth operation of applications and services. 

At its core, CloudWatch aggregates data in a detailed, programmatic manner, which enables users to gain insights, set alarms, visualize logs and metrics, troubleshoot issues, and discover patterns. This comprehensive coverage allows for the proactive management of the operational health of AWS resources such as Amazon EC2 instances, Amazon DynamoDB tables, and Amazon RDS DB instances, among others. One of the pivotal features of CloudWatch is its ability to monitor resource utilization, system-wide performance changes, and operational health in a unified manner. For example, it can track CPU usage, disk reads and writes, and network traffic across a fleet of EC2 instances. This capacity for close monitoring aids in the prevention of potential issues by setting alarms that notify users when thresholds are crossed or unexpected patterns emerge. 

Application Signals provides standardized application metrics and application-centric health views without requiring custom instrumentation. Its optional complete transaction-span visibility adds search and analytics beyond sampled transactions.

Container Insights with OpenTelemetry collects EKS metrics through open-source receivers and publishes them through the OTLP endpoint. It provides container-level CPU, memory, disk, network, and GPU visibility, with PromQL queries and automatic dashboards.

Classic Container Insights offers an optional enhanced-observability mode with container-level ECS and EKS metrics, Kubernetes state metrics, and EKS control-plane metrics. Without enhanced observability, it provides aggregated cluster- and service-level metrics.

Database Insights Standard Mode combines database metrics, SQL query analytics, and logs for Aurora and RDS. It includes up to seven days of detailed database and per-query metric history, with configurable retention from one to 24 months beyond that.

Cross-account observability lets a monitoring account view and query metrics and logs from source accounts without copying the data.

CloudWatch Synthetics runs customizable endpoint tests around the clock and alerts when behavior differs from expectations, covering checks such as availability, latency, transactions, and multi-step user workflows.

Real User Monitoring covers both web and mobile applications: web telemetry includes page-load delays and web vitals, while mobile telemetry uses OpenTelemetry spans and events.

X-Ray traces provide an end-to-end view of requests traveling through distributed applications, helping developers analyze and debug production behavior.

PromQL alarms have a non-adjustable quota of 15,000 per account per Region.

CloudWatch Omni provides an additional observability experience through an IDE extension for VS Code, Kiro, and Cursor and a standalone web UI for operations teams. Web users authenticate with enterprise SSO through IAM Identity Center and do not need AWS credentials, AWS Console access, or an AWS account.

Moreover, CloudWatch is not limited to predefined metrics and logs; it also allows users to publish their own custom metrics, offering flexibility that covers a wide range of use cases, from tracking application-level operations to specific business indicators. Through its dashboard, CloudWatch enables users to visualize these metrics in a comprehensible manner, facilitating easy digestion of data points and trends over time. This can be crucial for capacity planning, ensuring that applications and services are adequately provisioned to meet demand without overspending. 

Alongside classic CloudWatch metrics, CloudWatch natively ingests OpenTelemetry metrics and supports PromQL queries, allowing existing Prometheus dashboards and alerting workflows to use their existing queries.

Metric centralization creates physical metric copies in a destination account so alarms, dashboards, and queries can operate locally without cross-account calls.

A single cross-account metric stream can include metrics from multiple AWS accounts within one AWS Region.

In addition to monitoring, CloudWatch Logs provide a centralized repository for storing log information, including system logs, application logs, and any custom log data, with the ability to define filters and set alarms on log patterns. This capability is integral for troubleshooting, as it aids in pinpointing the root causes of operational problems quickly. CloudWatch Events (now part of Amazon EventBridge) further expands its capabilities by allowing AWS resources to automatically react to system-wide changes through a stream of real-time data from AWS services, transforming and responding dynamically to operational changes across the AWS landscape. 

Logs Insights supports log queries with aggregations, filters, and regular expressions, as well as time-series visualization, drill-down into individual events, and export of query results to CloudWatch dashboards.

Contributor Insights continuously analyzes time-series data to identify the top contributors influencing system performance.

To sum up, Amazon CloudWatch plays a critical role in ensuring that AWS-based applications run smoothly and efficiently. By providing detailed insights into applications and system operation, it not only helps in proactive issue detection and resolution but also aids in understanding and optimizing system performance, leading to more robust, reliable, and efficient application deployments. Its integration capabilities, coupled with the depth of insights it provides, make CloudWatch a cornerstone in managing cloud operations effectively.

Concepts related to Amazon CloudWatch

AWS architecture icons are the property of Amazon Web Services. BlowStack is not affiliated with or endorsed by AWS.