No		Q	A
I.1	✅	Improve performance of S3 upload	Multipart upload. How about Transfer Acceleration?
I.2	✅	Caching: Lazy-load & Write-through
I.3	✅	Lambda: Different parameter depends on environment	Environment variable (# Stage variable)
I.5	✅	API Gateway: same API Gateway for multiple environments	Stage variable
I.6	✅	CloudWatch Alarm: When to alarm?	Period / Evaluation Periods / Data points to Alarm
I.7	✅	SAM: Nested application	AWS::Serverless::Application
I.8	✅	Lambda deploy: Graduate increase traffic to new version	Lambda supports traffic shifting for aliases (by setting traffic weight). (That’s why SAM have canary, linear deployment strategy )
I.9	✅	API Gateway: Map data for microservice (in container)	Use HTTP integration (Not AWS or _PROXY)
I.10	✅	Database scale globally, handle frequent schema changes	DynamoDB (not Aurora)
I.11	✅	API Gateway - TTL 300s. How client can invalidate cache?	Send request with Cache-Control: max-age=0 header
I.13	✅	Session data store on DynamoDB. Delete session of logged out users?	Use DynamoDB TTL (when will an item is eligible for expiration - in epoch time)
I.15	✅	Serve content based on location	- CloudFront can add header about viewer location’s location (based on the viewer’s IP address).
I.16	✅	What to do after update Cfn template of a Lambda function?	1. aws cloudformation package: Uploads local artifacts to S3, update Cfn template to reference these artifacts
I.17	✅	Multi-thread key-value cache store	Elasticache for Memcached
I.19	✅	Process long-running tasks	Elastic Beanstalk worker environment (an HTTP request handler that EB invokes with an SQS queue)
I.20	✅	ECS: Where to config port for container?	Task definition
I.22	✅	Can Lambda handle 10 requests/s (each request take 50s)	500 concurrent execution (Lambda can handles without doing anything. The default limit is 1.000)
I.23	✅	DynamoDB table attributes: UserID (PK) - GameTitle (SK) - TopScore. Get max TopScore of each game?	Create a Global Secondary Index: GameTitle (PK) - TopScore (SK). Query 1 item for each game in descending order.
I.24	✅	Kinesis Data Streams: How many worker is optimal to process a number of shards?	1:1 ratio
I.25	✅	AWS CLI timeout when list bucket with 10.000 objects	Add pagination parameter when use AWS CLI
I.26	✅	Permission to Create/Delete GitCommit repos	codecommit:CreateRepository & codecommit:DeleteRepository
I.27	✅	DynamoDB Streams: Send a copy of old item to S3	StreamViewType: OLD_IMAGE
II.1	✅	RDS + Lambda + Translate text with Amazon Translate. Improve performance, reduce load to RDS?	Use Lambda execution environment /tmp as cache store
II.2	✅	DynamoDB: Employee info. Which attribute use as primary key?	employee_id
II.3	✅	Run a Lambda function every 30min?	Use EventBridge Schedule Rule to create scheduled events -> target the Lambda function
II.4	✅	ECS integrate with API Gateway. Which integration make no intervention?	HTTP_PROXY
II.5	✅	Which API to call to get permission (from a IAM Role) to interact with S3 API	AWS STS AssumeRole API (or AssumeRoleWithWebIdentity, AssumeRoleWithSAML) (optionally pass an SessionPolicy)
II.6	✅	Asynchronous invoke Lambda function with Invoke API	Pass parameter InvocationType - Event
II.11	✅	DynamoDB: Write heavy - ProvisionedThroughputExceededException. Why?	RCU/WCU of GSI is separate from base table
II.12	✅	CodeCommit: Permission to fetch, clone, push	codecommit:GitPull, codecommit:GitPush
II.13	✅	Elastic Beanstalk: Deploy new version with CLI	Package app as zip/war file. Deploy with eb deploy (EB CLI is a different package. It’s not packaged with AWS CLI, not aws eb)
II.14	✅	Internal app: elastic, cost-effective	DynamoDB, EC2 Spot Fleet
II.15	✅	Build an CI/CD to deploy to both EC2 & on-premise. Which service?	CodeDeploy
II.16	✅	DynamoDB: 10 RCU, each item 4KB. How much read request/s the table can handle?	10 RCU: 10 Strong Consistency Read - 20 Eventually Consistency Read
II.17	✅	Prototype microservices on ECS. Which task placement minimize cost?	Binpack, random, spread
II.18	✅	Temporary AWS credential for both authenticated/unauthenticated. Which service?	Cognito Identity Pool (Federated Identity)
II.20	✅	SQS: Duplicated messages. How to fix?	For SQS, when call SendMessage API, add MessageDeduplicationId param
II.21	✅	DynamoDB: Protect from overwritten?	Implement optimistic locking with version number (ensure update the right version item)
II.22	✅	Lambda: Asynchronous invoke, exponential back-off, then send un-processed messages to another service.	Lambda supports Dead Letter Queue (just as SQS DLQ)
II.23	✅	Collect visitor click to ads. Which service?	DynamoDB: Use UpdateItem to implement atomic counter (Approximately ~ Add 1 to previous value)
II.24	✅	DynamoDB: Read 100 items. Which API?	BatchGetItem (instead of GetItem) then BatchWriteItem
II.25	✅	DynamoDB: Store recent updated item automatically	1. DynamoDB Streams + Lambda (not recommend)
II.26	✅	DynamoDB: Session data. Reduce storage without using provision throughput	Use DynamoDB TTL
II.27	✅	Kinesis Data Streams: How to handle data flow (Scaling)?	1. Split shards to increase stream capacity; 2. Merge shards to decrease stream capacity.
II.28	✅	Kinesis Data Streams: Consumer process every other day & store to S3. S3 only has half of the data?	Kinesis Data Streams default retention period is 24h (up to 365 days)
II.29	✅	An AWS account has 2 Lambda function. 1 works optimal, 1 is throttled. Why?	The reserve concurrency execution of function 1 is higher
II.30	✅	Lambda: Process events from S3 Events. 10 events/s, each event takes 3s. How much concurrency?	Concurrency: number of in-flight request AWS Lambda function is handling at the same time.
II.32	✅	Protect AWS APIs call with MFA	MFA-enabled IAM users must call GetSessionToken and submit an MFA code that is associated with their MFA device.
II.34	✅	Website hosted on S3 bucket_A, make GET request to S3 bucket_B. API calls are blocked by browsers?	Enable CORS on bucket_B
II.35	✅	Migrate repos from Github to CodeCommit	1. Create repos on CodeCommit; 2. Clone repos from Github; 3. Push to CodeCommit repos
II.36	✅	Lambda: Increase CPU?	By increase memory
II.37	✅	ECS: Integrate with X-Ray. How to provide more granular timing information?	Use sub-segment
II.38	✅	Version control system. Which service?	CodeCommit
II.39	✅	API Gateway: Non-proxy integration with Lambda. How to ensure consumer include a query string?	Use method request
III.1	✅	Replace spreadsheet-based tracking system. DynamoDB.	EDA: DynamoDB Streams + Lambda + SNS
III.2	✅	Collect real-time user data	Kinesis Data Streams
III.4	✅	CodeCommit, CodeBuild, CodeDeploy, CodePipeline. A central dashboard?	CodeStar
III.5	✅	Cfn: Reuse value from a stack?	Outputs/Export & Fn::ImportValue
III.6	✅	Hundreds of thousands of reads/writes per second. Which database?	DynamoDB
III.7	✅	DynamoDB: Bidding system.	DynamoDB Streams + Lambda + Conditional Write
III.8	✅	DynamoDB: Query a single table	LSI (when create the table)
III.9	✅	DynamoDB: Throttled. Hot partition. Fix minimal effort?	Refactor to evenly distribute between partitions; Retries with exponential backoff
III.10	✅	CodeCommit: Setup?	1. Git credential - HTTPS (username/password); 2. SSH connection (SSH key-pair)
III.11	✅	DynamoDB: 1.5KB. Write 100 items/s. WCU 100 but throttled. Fix?	Increase WCU to 200.
III.13	✅	SQS: Duplicate message	SQS FIFO + SendMessage with DeduplicationID
III.15	✅	ECS: Place task between AZs?	Spread
III.17	✅	Sync user data without your backend.	Cognito Sync (delegated use AppSync)
III.21	✅	SQS: Config Dead Letter Queue	Just provide ARN of the queue to DeadLetterConfig
III.22	✅	Elastic Beanstalk: Web app process large number of items from DynamoDB. Overload. Easiest fix?	Use Batch operations for Get, Put, Delete
III.23	✅	DynamoDB: Send welcome mail for new user. How?	DynamoDB Streams + Lambda + SNS
III.24	✅	DynamoDB: Table - FighterID (PK) - FilterTitle (SK). Query by other attributes?	Create a GSI
III.25	✅	DynamoDB: + Elasticache. Write data if cache miss. Improve?	Add write-through + TTL
III.26	✅	Kinesis Data Streams: 100 shards, Lambda (10 seconds/request, 50 items/seconds)	Maximum of 100 Lambda concurrency = number of shards
III.27	✅	Cfn: How to automate the process of getting latest AMI?	Use System Managers Parameter Store
III.28	✅	SQS: Process tool long & messages appear twice	Increase visibility timeout
III.29	✅	CloudFront: Slow to login & 504	Authentication@Edge + Origin fail over
III.30	✅	Lambda: Increase CPU?	By increasing memory
III.31	✅	API Gateway: Implement APIs form current Swagger spec.	Just import the OpenAPI/Swagger file.
III.32	✅	Microservice using Docker + Fine-grain control	ECS
III.33	✅	Quickly deploy Node app (provisioning, load balancing, ASG…)	Elastic Beanstalk
III.34	✅	Amplify: config?	amplify.yaml
III.35	✅	S3: Cross-Region Replication. Fail. Why?	Maybe Object Versioning is not enable
III.36	✅	S3: Upload Terabytes of data from over the worlds. Slow. Improve speed?	S3 Transfer Acceleration
III.37	✅	API Gateway: Reuse same function for different stage (different DynamoDB table)	Stage variable + mapping template
IV.1	✅	Automatically watermark images uploaded to S3	Use S3 Event Notification s3:ObjectCreate:Put. Send the event to Lambda function (destination) (# S3 Object Lambda)
IV.2	✅	Lambda function: access RDS in private subnet	Connect to the VPC
IV.3	✅	API Gateway + Lambda: Map incoming request to integration request & vice versa	AWS custom (Lambda custom)
IV.4	✅	Which cache strategy ensured cached data is always up-to-date & stale data are automatically deleted?	Write-through + With TTL
IV.5	✅	DynamoDD: How to prevent data overwritten?	Optimistic locking + Conditional writes (check version number)
IV.6	✅	Database credential: How to encrypted & auto-rotate?	Use Secrets Manager + Enable Automatic Rotation
IV.7	✅	S3: Use SQL to filter & retrieve only a subset data of an object?	S3 Select (supports CSV, JSON, Apache Parquet & compressed CSV, JSON)
IV.9	✅	DynamoDB - GSI: How much throughput to provision?	GSI WCU >= Base table WCU
IV.10	✅	DynamoDB: 2KB items - 10 writes/s - 20 eventually consistent reads/s. How much RCU/WCU?	2KB ➡️ 1 strongly consistent read = 1 RCU ➡️ 1 eventually = 0.5 RCU ➡️ 20 eventually = 10 RCU
IV.11	✅	1 item = 17KB. 320 strongly consistent read/s. How much RCU?	17KB/item ➡️ 5RCU/strongly consistent read ➡️ 320 strongly read = 320 x 5 = 1600 RCU
IV.12	✅	HTML, JS, image, video. How to server with lowest latency around the world?	S3 + CloudFront
IV.13	✅	AI-based app built with Lambda. How to modify the way invocation event are read form Lambda runtime API?	Use Lambda custom runtime
IV.14	✅	Migrate monolith on-premise app to Lambda. Best practice?	1. Take advantage of execution runtime; 2. Use environment variable…
IV.16	✅	Online game - Sync app pref + state of player + Allow multiple player to share state	App Sync (Cognito Sync not support shared data)
IV.17	✅	AWS CLI in EC2 instance. How to easily switch role?	1. Create a new CLI profile with credential; 2. Run aws CLI with --profile
IV.18	✅	DynamoDB: Ensure item is updated only some attribute meets some condition	Conditional writes
IV.19	✅	DynamoDB: Debug throughput of both base table & GSI when update item?	Call UpdateItem with ReturnConsumedCapacity set to None / Total / Indexes
IV.20	✅	DynamoDB Streams: How to integrate with Lambda?	1. Create an EventSourceMapping to poll the DynamoDB stream, read & process records
IV.21	✅	Serverless app: Which service can manage configuration & deploy the whole stack + simple?	AWS SAM (Serverless Application Model)
IV.22	✅	DynamoDB: Table: ArticleName (PK) - Category (SK). Query ArticleName using another Sort Key + Strongly consistent read	Create a new table with Local Secondary Index (LSI). Migrate the existing data to new table (LSI cannot be created after the DynamoTB is created)
IV.23	✅	S3: Ensure all objects are encrypted with AE256	Use bucket policy to deny any Create request doesn’t have x-amz-server-side-encryption: AE256 header
IV.24	✅	EC2: Shell script to get instance public/private IP	Use Instance Metadata Service endpoint 169.254.169.254/latest/meta-data
IV.25	✅	Lambda: Account concurrency limit 2000; 10 functions: 1 function 400, 1 function 200. The rest, the third?	The rest: 1400; the third: 1300.
IV.26	✅	Coordinate multiple services into serverless workflow. Which service?	AWS Step Functions
IV.27	✅	API Gateway: Enable caching. How to invalidate 1 key?	Send the request with Cache-Control: max-age=0
IV.28	✅	Lambda: Connection refused. Why?	Maybe the invoke URL is wrong (http without s)
IV.29	✅	Lambda: Improve performance?	Increase memory will increase CPU too.
IV.30	✅	Lambda: A function initialize DB connection every time it executes. How to optimize?	Move the DB connection to shared execution context (outside handler)
IV.31	✅	Lambda: Can the function built with Rust?	Yes. Use custom runtime
IV.32	✅	Lambda: A function fetch 20MB static data every time it executes. How to optimize?	Place the initialize outside Lambda handler; saved external file to /tmp
IV.34	✅	Online game. How to add feature to cross-sync profile data between device?	Use Cognito Sync (or App Sync)
IV.35	✅	DynamoDB: Which attribute to use as partition key?	The partition should uniquely identify each item
IV.36	✅	CodePipeline: Push build details into a DynamoDB?	Use EventBridge & Lambda
V.2	✅	API Gateway & Lambda: New version. Smooth migration	Update Lambda function -> Deploy new version. Specify new ARN in API Gateway integration. Redeploy to new stage.
V.3	✅	Elastic Beanstalk: Path of config file	/.ebextensions/xxx.config
V.4	✅	Lambda: Ephemeral storage	/tmp
V.5	✅	Lambda: How to use Ruby?	Ruby’s natively support
V.6	✅	CloudFront: Update image immediately	Use file name versioning
V.7	✅	Cognito: Authentication with JWT	1. Create User Pools
V.9	✅	S3: Lowest cost	Glacier Deep Archive
V.10	✅	EC2: Run Apache web server	Use user-data to install and start Apache web server
V.11	✅	App host in 1 region. Re-create on other regions using AMI & CloudFormation. How?	Use Cfn Mapping & FindInMap
V.12	✅	Serverless app. Application code & infrastructure code in Python. How?	Use CDK and Python
V.13	✅	Step Functions: Handle error? Aggregate data in different states?	Use Catch & ResultPath
V.15	✅	DynamoDB: Control access to individual items & attributes?	Fine-grain access control with
V.16	✅	Cognito: UI for login page missing brand logo	Cognito allows customization for: logo, CSS.
V.17	✅	SQS queue: messages larger than 256KB.	Use Amazon S3, (for Java) Extended Client Library
V.20	✅	API Gateway: Integrate a XML-based SOAP API. How?	Use HTTP Integration:
V.21	✅	Serverless: Send newsletter at 7-day interval. How?	Use EventBridge Schedule Rule to create events at 7-day interval, sends the events to Lambda function target
V.23	✅	DynamoDB: Group multiple actions to multiple items to a one-or-nothing operation?	Use TransactionWriteItems
V.24	✅	DynamoDB: When an item added to Customer table, dynamically update Payment table. How do it in real time?	Enable DynamoDB Streams for Customer table, trigger a Lambda function to update Payment table
V.25	✅	ALB: Obtain all value of identical query parameter key.	Enable multi-value headers
V.26	✅	ECS: 2 containers share logs. How?	Define these 2 containers in 1 task definition, use EFS as a volume

No		Q	A
I.2	✅	Share DB endpoint	Use System Manager Parameter Store secure string
I.4	✅	API Gateway: Allow another account invoke via IAM Role	1. From the other account, grant permission to interact with this API Gateway
I.5	✅	Grant permission to access only some path of a S3 bucket	Use S3 bucket policy, with the policy statement Resource including the path
I.6	✅	Login with social site: Facebook…, then access to AWS service	Cognito Identity Pool (Federated Identity)
I.7	✅	S3 - Encrypt AE256	x-amz-server-side-encryption: AE256
I.9	✅	KMS - Generate data key but not use immediately	GenerateDataKeyWithoutPlaintext
II.1	✅	Cognito - Enforce MFA for suspicious login attempt	User Pool / Adaptive Authentication / Automatic risk response
II.2	✅	EC2 instance - Upload images to S3. How to give EC2 instance permission?	IAM Role (Instance Profile)
II.3	✅	Dev needs to access Test/Prod accounts. How to give permission?	Grant cross-account access:
II.4	✅	API Gateway - Lambda Authorizer: Implement one that same as OAuth, SAML	API Gateway - Lambda Authorizer (Custom Authorizer) has 2 types:
II.5	✅	EC2 instance has both credentials & Instance Profile	AWS CLI credentials precedence: CLI options > ENVVAR > Assumed Role > IAM Identity (/.aws/config) > Credential file (/.aws/credentials) > EC2 instance profile > Container credential
II.8	✅	S3 - Encryption using KMS-C. How to upload object?	Include the headers: x-amz-server-side​-encryption​-customer- + algorithm/key/key-MD5
II.9	✅	IAM, SCP - How to test IAM policy?	IAM Policy Simulator
III.1	✅	API Gateway: Lambda authorizer - accepts header, query strings	Request parameter-based authorizer (REQUEST authorizer)
III.2	✅	On-premise KMS, migrate to AWS. Key is store in dedicated hardware	CloudHSM
III.3	✅	S3: Upload small file + Use KMS key: OK. Upload big file (100GB): not OK	AWS CLI use multipart upload for big files. It’s required the kms:Decrypt permission
III.4	✅	S3: SSE with KMS (using default KMS key). Which header?	- x-amz-server-side-encryption: aws:kms
III.5	✅	S3: SSE-C. How does it work?	1. You manage key & give AWS the key each time you upload a file
III.6	✅	KMS: A file encrypted with data encryption key (DEK). How to decrypt the files locally?	1. Use KMS’s Decrypt to decrypt the DEK
III.7	✅	CodeDeploy: Platform: ECS - appspec.yaml	appspec.yaml for ECS needs: TaskDefinition, ContainerImage, ContainerPort
III.8	✅	Share DB connection endpoint	Systems Manager Parameter Store secure string
III.9	✅	SSL certificate from 3rd party. Which service can store?	AWS Certificate Manager (for unsupported regions, use IAM certificate store)
IV.2	✅	SSM Parameter Store: Notify if param haven’t been rotated for 90 days	1. Use Advanced tier / Parameter polices / Notification policies
IV.3	✅	S3: Ensure data is encrypted at rest using the company key	1. Client-Side Encryption (CSE) - Encrypt the data before send to S3
IV.4	✅	KMS: Locally encrypt data (Envelope encryption)	1. Use the GenerateDataKey API to get a data key.
IV.5	✅	API Gateway: Custom authorizer using bearer token (same as SAML, OAuth). How?	Use API Gateway Lambda authorizer (aka custom authorizer)
IV.6	✅	Database credential + Rotate	Secrets Manager + Auto Rotation (How? Secrets Manager run a Lambda function )
IV.7	✅	CloudFormation: Retrieve license key + cost-effective	Systems Manager Parameter Store + Secure String
IV.8	✅	Migrate on-premise to AWS. Integrate LDAP directory service (not compatible with SAML)	Implement a custom identity broker, which use STS to issue short-live AWS credentials
IV.9	✅	Cognito: Additional authentication method	Integrate Multi-Factor Authentication (MFA) to Cognito User Pool
IV.10	✅	SQL Server. Migrate to RDS. Encrypt data before write to disk & vice versa.	Enable SQL Server Transparent Data Encryption
IV.11	✅	S3. Someone use your image without permission.	(Block public access) Use pre-signed URL / pre-signed cookies
V.2	✅	S3 bucket in production account. How to allow a user on dev account access?	1. In prod acc, create an IAM role
V.3	✅	S3. Which service to allow user register/sign-in & upload/access images on S3.	Cognito User Pools & Identity Pools
V.4	✅	Allow temporary access to EC2 & but still enforce MFA? Which STS API?	STS GetSessionToken
V.6	✅	API Gateway: Regulate access to API & charge based on usage	Usage Plan
V.7	✅	Best practice to manage access key	Remote all access key of root account, use IAM role for applications
V.8	✅	Most secure way to send CW logs in EC2 instance of ASG launch configuration	- Create a new IAM role for the new Launch Configuration
V.10	✅	S3: Hundreds of thousands of objects. Turn on SSE-KMS. Performance degradation. Why?	Requests to KMS are exceeded quota
V.11	✅	How to check permission of an IAM role?	1. Use IAM Policy Simulator

No		Q	A
I.3	✅	Deploy Lambda with CodeDeploy	CodeDeploy deploy configuration:
I.4	✅	Build, test, deploy serverless app	Serverless Application Model (SAM)
II.1	✅	SAM: workflow redeploy	1. sam build; 2. sam deploy
II.2	✅	Elastic Beanstalk: Dev/Test - Deploy ASAP	EB deployment strategy: AllAtOnce
II.3	✅	SAM: deploy 10% every 10min	SAM Linear 10min
II.4	✅	Lambda: package deployment take too much time to deploy? How increase deploy speed?	Lambda package deployment quota (for a function) is 50 MB (zipped; includes runtimes, layer…)
II.6	✅	CloudFormation: Easiest way to deploy a hello_world lambda function	Incline the function in AWS::Lambda::Function Properties / Code / Zipfile
II.7	✅	Elastic Beanstalk: Upgrade from Java 7 to 8. Shift all traffic to the new one, revert if something’s wrong	Blue/Green Deployment: EB AllAtOnce + EB swap environment URLs (EB will use Route 53 to swap CNAMEs)
II.8	✅	CodeDeploy: Deploy to where? How?	CodeDeploy deploy configuration:
III.1	✅	EC2: EBS-backed root volume. How to detach the root-volume?	Stop the EC2 instance, then detach the root volume
III.2	✅	Elastic Beanstalk: Maintain compute resource while deploying. No downtime.	Rolling with additional batch, Immutable
III.3	✅	DynamoDB: 3.5KB. 150 eventually consistent reads/second. How many RCU?	1 strongly read -> 1 RCU
III.4	✅	Lambda: Deployment package 80MB. What to do now?	Split the dependencies to a layer
III.5	✅	SAM: How to use Cfn & include SAM?	Use Cfn Transform & AWS:Serverless macro to process SAM template to Cfn template
III.7	✅	Cfn: Different accounts. How to manage update across all accounts?	Use StackSets
IV.2	✅	Elastic Beanstalk: HA, revert quickly?	1. Use any EB deployment strategy, e.g. All-at-once (fastest)
IV.3	✅	CloudFormation: Inline code in template	AWS::Lambda::Function / Code / ZipFile
IV.4	✅	Serverless app: Zip code, upload to S3, produce package deployment-ready template & deploy	sam deploy (which includes sam package)
IV.5	✅	CloudFormation: Install packages, start services on EC2 after provisioned	cfn-init
V.1	✅	CodeDeploy: S3 source. Deploy to EC2 fail during DownloadBundle	The EC2 instance profile don’t have permission to access S3 bucket
V.3	✅	Deploy to 3 environments: test, staging, production. How?	Use 3 deployment groups, each for 1 environment.
V.4	✅	Elastic Beanstalk: Multi-container Docker. Which file to configure container definitions?	Dockerrun.aws.json
V.5	✅	Elastic Beanstalk: Multi developers deploy without upload the whole project?	Use eb deploy to deploy from local CodeCommit repo
V.6	✅	Serverless app defined with CDK. Deploy to new account. NoSuchBucket error. Fix?	Run cdk bootstrap to provision resources for CDK deployment, e.g. IAM role, S3 bucket.
V.8	✅	CodeCommit: Forgot to pull master. Fix conflict?	git rebase feature branch on master branch. Then manually fix conflict.

No		Q	A
I.1	✅	Move session data to AWS, 100% CPU, HA	ElastiCache for Redis
I.2	✅	sam local - Access denied	1. aws configure --profile; 2. sam local invoke --profile
I.3	✅	Cost effective when using SQS	Long polling (config ReceiveMessage WaitTime)
I.4	✅	Record all changes to a DynamoDB table to another one	DynamoDB Streams + Lambda (write to another DynamoDB table)
I.5	✅	DynamoDB - Reduce queries	DAX
I.8	✅	ECS store data in DynamoDB, how to verify each new items?	DynamoDB Streams + Lambda
I.9	✅	API Gateway - 504	504 Gateway timeout -> Integration timeout after 30s
I.10	✅	RDS - Too many connections	RDS Proxy
I.11	✅	API Gateway timeout - Which CW metrics to watch?	Latency, IntegrationLatency
I.13	✅	CloudFront - end-to-end SSL	Viewer Protocol Policy, Origin Protocol Policy: HTTPS
I.15	✅	Capture IP in/out of an VPC	VPC Flow Log
I.16	✅	Elastic Beanstalk config for cron-job	cron.yaml
I.17	✅	DynamoDB - Fetch only some attributes	Use projection attributes
I.18	✅	Lambda function download same big file	Use /tmp
I.21	✅	App deploy with Elastic Beanstalk - Config X-Ray	.ebextensions/xray-daemon.config
II.1	✅	DynamoDB: ProvisionedThroughputExceeded	Optimize DynamoDB:
II.2	✅	X-Ray: View full trace without using console	Use GetTraceSummaries, BatchGetTraces
II.3	✅	ECS + Docker app: How to setup tracing with X-Ray?	- Run xray-daemon inside docker image
II.4	✅	AWS CLI - Timeout	Use pagination: --page-size (request less items) --max-items (show less items)
II.9	✅	X-Ray: From where, IP address is fetch?	X-Forwarded-For header
III.2	✅	X-Ray: Insufficient permissions to use X-ray console to view service map, segments. Which manged policy?	- AWSXrayReadOnlyAccess
III.3	✅	API Gateway: Fetch latest data without caching (using Cache-Control: max-age=0 header). Which permission?	Resource-based policy for execute-api:InvalidateCache action
III.4	✅	DynamoDB: Streams, EventBridge + Lambda every 36 hours. Missing data?	DynamoDB Streams retention period is 24 hour. Only last 24 hours data is available
III.5	✅	Kinesis Data Streams: Increasing data flow. Scale up?	Split-shard + (increase numbers of worker)
III.6	✅	S3: CORS config: <AllowOrigin>, <AllowedMethod>, <AllowedHeader>, <MaxAgeSeconds>3600</MaxAgeSeconds>	MaxAgeSeconds: time in seconds that your browser can cache the response for a preflight request
III.7	✅	API Gateway: Lambda. 504. No errors in CW. Why?	Lambda function takes more than 30s (API Gateway timeout)
III.8	✅	RDS: Slow response (in peak time). Already optimize queries. Resolve?	- Add Read Replica
III.9	✅	Latency-sensitive service. AWS Fargate, CloudFront, ALB. Too much unauthenticated users, increase CPU of Fargate. Fix?	Use CloudFront Function (attach to Viewer Request) to authenticate users
III.10	✅	EC2: Monitor memory, swap. How?	Install CW Agent
III.11	✅	Elastic Beanstalk: EC2. CW doesn’t show memory. Why?	By default, CW doesn’t track EC2 instance memory
III.12	✅	Kinesis Data Steams: Producers restart -> Duplicate record. Fix?	Call PutRecord with SequenceNumberForOrdering param.
IV.1	✅	Send traces to X-ray?	Use X-Ray daemon (CloudWatch Agent can do this on EC2/On-Premise)
IV.2	✅	SQS: Duplicate message. Fix?	Use FIFO queue + provide deduplicationID
IV.3	✅	X-Ray: Filter trace?	1. Use web console; 2. Use GetTraceSummaries (support search)
IV.4	✅	X-Ray: Send trace to X-Ray?	- 1. Use X-Ray SDK (through X-Ray daemon); 2. Use X-Ray/CLI (directly)
IV.5	✅	Kinesis: Over-provision. Scale in?	Merge cold shards
IV.6	✅	API Gateway: Terminated Lambda. Why?	- Lambda timeout: max 15min ➡️ terminated
IV.7	✅	API Gateway: No metrics for CacheHitCount/CacheMissCount	API Gateway caching is not enabled
IV.8	✅	SQS injections, XSS attack. How to deal?	Use Web Application Firewall (WAF). It works with: CloudFront, ALB, API Gateway (REST API)
IV.9	✅	Kinesis: 10 shards - 10 EC2 instance. Increase to 20 shards, how many instances?	20 instances, the number of instances match the number of shards by 1:1 ratio
IV.10	✅	RDS: Monitor memory, CPU usages of processes?	Use RDS enhanced monitoring
IV.11	✅	CodePipeline: Code review in each stage before move to next stage	Use a “manually approval” action, and send the approval request to a SNS topic
IV.12	✅	X-Ray: Record call to DB, other services, SQL queries & filter	Add annotations in the subsegment section of the segment document.
IV.13	✅	X-Ray: Permission to send trace to X-Ray?	AWSXRayDaemonWriteAccess
V.3	✅	CodeBuild: Run on a proxy server. RequestError when CodeBuild is accessed. Fix?
V.5	✅	S3 - Event Notifications: Compress the images, but it takes too much time. Improve?	Increase memory ➡️ increase CPU
V.6	✅	Debug latency of your app (with recently added function)? How to do with X-Ray?	Define sug-segments inside the function to “instrument” (measure) it
V.7	✅	AWS CLI: Create snapshot of EC2 instance. InvalidInstanceID.NotFound. Why?	Maybe the AWS CLI is using a profile for a different region
V.8	✅	Build a CI pipeline. Which AWS services?	CodeCommit, Lambda, CodeBuild
V.9	✅	Step Functions: Handle & recover from State’s exception.	Use Catch & Retry fields in state machine definition
V.10	✅	Lambda - Cold start: Optimize?	1. Reduce pre-handler code
V.11	✅	LAMP stack. Migrate to AWS?	EC2 + Aurora/RDS
V.12	✅	Lambda: process file (5min). So slow?	Change InvokeType to Event (asynchronous invocation)
V.14	✅	Website (hosted on S3) call API Gateway. No "Access-Control-Allow-Origin" error. Fix?	Config CORS for API Gateway to allow the website (S3)
V.15	✅	API Gateway + Lambda: Publish a new version of AccService:Prod with the alias AccService:Beta. How to test before promote?	Create a BETA stage. Use stage variable to reference the beta function alias
V.16	✅	Lambda: Unable to import module. Fix?	1. Install the missing module locally.
V.17	✅	Elastic Beanstalk: Keep the old code in S3 bucket. How?	Change Retention to Retain source bundle in S3