AWS Unpacked #15: More AWS Tools Worth Knowing

Elastic Beanstalk

Elastic Beanstalk is a Platform-as-a-Service (PaaS) offering that lets you deploy and manage applications without worrying about the underlying infrastructure. You just upload your application code, and Beanstalk handles provisioning, load balancing, scaling, and monitoring.

Key Concepts:

• Application: A logical collection of environments.
• Environment: The infrastructure where your app runs. Can be updated, cloned, or deleted independently.
• Tiers:
  • Web Server Tier: Handles HTTP(S) requests, typically for frontend apps.
  • Worker Tier: Handles background jobs using Amazon SQS.
• Deployment Modes: All at once, rolling, rolling with additional batch, immutable—these control how updates are rolled out.
• Supported Platforms: Node.js, Python, PHP, Java, Ruby, .NET, Go, Docker—even multi-container Docker and custom platforms.

Use Cases:

• Deploying a web app with minimal infrastructure knowledge.
• Running a queue-worker job system with the worker tier.
• Rapid prototyping where time-to-market is critical.

Beanstalk is a higher-level abstraction over EC2, Auto Scaling, and ELB. If you need more control, go with ECS or EC2.

AWS CloudFormation

CloudFormation is a declarative infrastructure-as-code service. You define resources in a YAML or JSON template, and AWS provisions them in the correct order with the exact configuration that you specify.

Key Features:

• Supports nearly all AWS resources (plus custom resources).
• Makes your infrastructure reproducible and version-controlled.
• Tags resources so you can track and forecast costs by stack.
• You can estimate pricing before deployment.
• Enables DevOps workflows like spinning down non-prod stacks overnight for cost savings.

Benefits:

• Productivity: Easily destroy and recreate infrastructure.
• Savings: In Dev, you could automate deletion of templates at 5 PM and recreated them at 8 AM, safely
• Collaboration: Review infra changes through code (like Git).
• You can visualize and build templates using Application Composer, which helps map out architectures like WordPress stacks with clear component relationships.

Security:

• Use service roles to grant CloudFormation permissions to create/update resources, while users only need permission to call CloudFormation itself. This gives users the ability to create/update/delete the stack resources even if they don’t have permission to work with the resources in the stack. The user must have the iam:PassRole permission.

Analogy: Think of CloudFormation as a recipe (template) for your cloud “kitchen.” It tells AWS exactly how to cook up your infrastructure.

How does CloudFormation and Elastic Beanstalk compare?

Amazon SES (Simple Email Service)

A fully managed service for sending transactional, marketing, and bulk emails. You can send messages via the console, SMTP, or API. Also allows for inbound emails.

Features:

• Supports DKIM and SPF for email authentication.
• Choose between shared, dedicated, or customer-owned IPs.
• Real-time stats: deliveries, opens, bounces, complaints.
• Integrates with CloudWatch for performance monitoring.
• Flexible IP deployment: shared, dedicated, and customer-owned IPs

Use SES when your app or service needs to send out email reliably at scale—for example, confirmation emails, password resets, or newsletters and marketing/bulk email communications

Compared to Amazon Pinpoint, SES is lower-level: you handle recipients, content, and scheduling yourself.

Amazon Pinpoint

A multi-channel marketing communications platform for sending personalized, high-volume messages across SMS, email, push, and voice.

Highlights:

• Two-way communication (e.g., SMS replies).
• Build campaigns with segmentation, templates, and delivery schedules.
• Scales to billions of messages per day.
• Streams events (delivery success, failures) to Kinesis or CloudWatch.

Pinpoint is ideal for campaign-based messaging. If you’re running a customer engagement strategy (vs just sending notifications), use Pinpoint. SES and SNS are better for one-off or system-driven messaging.

Systems Manager (SSM) – Session Manager

Session Manager lets you remotely access EC2 instances via the AWS console or CLI - no SSH, no bastion host, no port 22.

Advantages:

• More secure than traditional SSH (no open ports or key management).
• Works with IAM permissions and integrates with CloudTrail for audit logs.
• Supports Windows, Linux, and macOS.
• Can log sessions to S3 or CloudWatch.
• Execute a script or just run a command
• Send notifications to SNS about command status

Great for highly secure environments where you want to avoid exposing EC2 instances publicly.

SSM – Other Services

Run Command:

• Execute scripts or commands across multiple EC2 instances without SSH.
• Useful for applying updates or running batch scripts.
• Output can be sent to S3 or CloudWatch, with status alerts via SNS.
• Can be triggered via EventBridge.

Patch Manager:

• Automates patching of OS and apps on managed instances.
• Generates compliance reports.
• Supports maintenance windows for scheduled patching.

Maintenance Windows:

• Define time blocks for running tasks like updates or reboots.
• Helps avoid downtime during business hours.

Automation:

• Runbooks (SSM Documents) define common maintenance workflows.
• Use to reboot instances, create AMIs, snapshot EBS, etc.
• Trigger via Console, CLI, EventBridge, or AWS Config.

Together, these SSM tools allow for secure, automated, centralized management of your EC2 fleet—especially at scale or in hybrid environments.

AWS Cost Explorer

Provides interactive dashboards and reporting for your AWS spend.

Capabilities:

• Visualize costs and usage over time.
• Break down usage by account, service, tag, etc.
• Forecast future usage up to 12 months based on trends.
• Analyze cost data hourly or at the resource level.
• Simulate Savings Plans and optimize purchasing decisions.

Use it for chargeback, budgeting, or just to make sure you’re not spending unnecessarily.

AWS Cost Anomaly Detection

Monitors your spend using machine learning to detect anomalies in real time.

Features:

• No need to set thresholds—it learns your patterns.
• Alerts via SNS or email.
• Pinpoints root causes and affected services/accounts.
• Helps avoid surprise bills (e.g., accidentally leaving something running).

This is a good complement to Cost Explorer, especially for accounts with unpredictable or spiky usage.

AWS Outposts

Outposts extend AWS infrastructure, APIs, and services to on-premises data centers.

• For workloads requiring low latency, local data processing, or data residency.
• Fully managed by AWS but physically installed in your data center.
• Supports EC2, ECS, EKS, RDS, S3, etc. on-prem.

Think of it as running AWS in your own building. Good for healthcare, finance, or industrial workloads that need to stay close to users or data sources.

AWS Batch

A fully managed service to run batch computing jobs at scale.

Key Points:

• Dynamically provisions EC2 (including Spot) to run your jobs.
• Jobs run as Docker containers, orchestrated using ECS.
• Ideal for jobs with a start and end (vs long-running services).

Use Cases:

• Image processing, scientific computing, big data ETL pipelines.

Batch vs Lambda: Lambda has runtime, memory, and duration limits. Batch is better for heavyweight or long-running jobs.

Amazon AppFlow

AppFlow lets you transfer data between SaaS apps and AWS securely and easily. Sources supported: Salesforce, SAP, Zendesk, Slack, ServiceNow

Examples:

• Pull Salesforce data into S3 or Redshift.
• Send Zendesk tickets into Snowflake.

Features:

• Event-driven, scheduled, or manual flows.
• Filter, transform, and validate data.
• Can run over AWS PrivateLink (bypassing the internet) or encrypted over the public internet

AppFlow saves you from building custom ETL pipelines or scraping APIs - it’s like Zapier, but AWS-native and enterprise-grade.

AWS Amplify

Amplify is a set of tools for building and deploying full-stack web and mobile apps.

Capabilities:

• Authentication, storage, APIs (REST/GraphQL), hosting, analytics, AI/ML.
• Git integration for CI/CD.
• Compatible with frameworks like React, Vue, Angular, Flutter.

It’s ideal for frontend and mobile developers who want to build serverless apps without worrying about backend infrastructure.

AWS Instance Scheduler

A solution to automatically start and stop EC2 and RDS instances based on custom schedules.

• Helps save money by shutting down non-prod environments after hours.
• Works by tagging resources and defining schedules.
• Often deployed via CloudFormation using pre-built templates.

While you could script this manually, Instance Scheduler makes it easy to automate across multiple services and accounts.

Machine Learning Services on AWS

Fully Managed ML Platform

Amazon SageMaker

A fully managed platform that lets developers and data scientists build, train, and deploy ML models at scale. SageMaker covers the full ML lifecycle—data prep, model training, tuning, deployment, monitoring—without needing to manage infrastructure. It supports Jupyter notebooks, built-in algorithms, AutoML, and integrates with most AWS services.

Computer Vision

Amazon Rekognition

Enables image and video analysis using ML. It can detect objects, people, scenes, and text, perform facial recognition and analysis, and track movements. Common use cases include:

• Face detection, search, and verification
• Celebrity recognition
• Labeling and scene detection
• Content moderation for detecting inappropriate material
• Text detection in images and video

Amazon Rekognition + A2I

When content moderation flags something with low confidence, results can be routed to Amazon Augmented AI (A2I) for human review.

Speech and Text

Amazon Transcribe

Automatic speech-to-text conversion. It uses deep learning to convert audio into text accurately, and can:

• Redact sensitive information (PII)
• Identify multiple speakers
• Handle multi-language inputs
• Support subtitles and closed captioning

Amazon Polly

Text-to-speech service that generates natural-sounding voice output. Supports multiple languages, dialects, SSML for voice customization, and even speech styles like “Newscaster.”

Amazon Translate

Neural machine translation for accurate, natural-sounding language translation. Ideal for websites, apps, and documents that need localization or multilingual support.

Amazon Lex

Lets you build voice and text-based conversational interfaces (chatbots) using the same tech that powers Alexa. Combines speech recognition and intent understanding. Can integrate with AWS Lambda for custom workflows.

Amazon Connect

A cloud-based contact center solution that integrates seamlessly with Lex. Enables natural conversations with customers via phone, supports CRM integrations, and scales without traditional hardware.

Natural Language Processing (NLP)

Amazon Comprehend

Analyzes unstructured text to uncover meaning and context. It can:

• Detect the dominant language
• Identify key phrases, entities, events, and sentiment
• Group documents by topics
• Extract syntactic structure

Amazon Comprehend Medical

A specialized version of Comprehend for the healthcare industry. It extracts information like medical conditions, medications, and treatment details from clinical text. It also detects PHI for compliance.

Recommendations and Forecasting

Amazon Personalize

Provides real-time personalized recommendations (e.g., product suggestions, marketing content). It’s based on the same tech Amazon.com uses. You feed in user behavior data, and it gives tailored results without needing to build ML models yourself.

Amazon Forecast

Time-series forecasting service that uses ML to generate highly accurate predictions for things like sales, inventory, or resource planning. Better than traditional statistical models, and fully managed.

Document Intelligence

Amazon Textract

Extracts structured data (tables, forms), handwriting, and text from scanned documents using ML. It works with PDFs, images, and more, and is especially useful in industries like finance, healthcare, and government for digitizing paper records.

Amazon Kendra

An ML-powered search service that lets you search across documents in natural language. Kendra can pull answers from PDFs, Word docs, web pages, etc., and gets smarter over time via user feedback.

Other Notable ML Tools

Amazon A2I (Augmented AI)

Allows you to incorporate human review into ML predictions, which is especially useful when decisions need extra oversight—like moderating content or verifying identity.

AWS Deep Learning AMIs

Pre-configured Amazon Machine Images with popular deep learning frameworks like TensorFlow, PyTorch, and MXNet. Great for running custom ML models on EC2.

AWS Inferentia

A custom chip designed by AWS for ML inference workloads. You can use it with SageMaker or EC2 to reduce cost and latency when deploying models at scale.