Skip to content

A. API-gateway - Stage 📚

  • url has API ID and stage name

1. Intro

img.png - once created API-g, need to deploy it on stages - give stage a name -dev,qa,prod,whatever - stages has deployment history - stage can be rollback. - so think of deploymnet object in k8s - has stage variable, for dynamic configuration for stages. eg: - lambda-arn - http-endpoint-uri - ... - note: passed to context object. => lambda.handler(event,context) 👈 - format ${stageVariables.var-1}

2. example

    - api-gateway-1:
      - integrated with lambda-1 arn ==> xxxx:${stageVariables.`lambda-alias`} 
        - created one stage variable
        - similarly create more stage variable.
      - set value of stageVariable "lambda-alias" in all stages, then
      - stages:
        - dev-1
          - set lambda-alias: dev
          - dev is point to $LATEST
        - qa-1
          - set lambda-alias: v10

3. cache setting

  • cache request/response
  • define at stage level
  • provision size : 0.5 GB to 237 GB
  • set cache object TTL
  • default: 5 min / 300 sec
  • max: 1 hr / 3600 sec
  • optionally encrypt
  • img_2.png

cache Invalidation 👈

  • from console
  • any client without/with iam:permission can do it
  • header: cache-control:max-age = 0
  • iam permission : execute-api:InvalidateCache

4. canary deployment

img.png - api-gateway-1 - stage-1 (95 % traffic) : monitor-1 - stage-2 ( 5 % traffic) : monitor-2 - console >> canary tab - set above configuration.

5. supports OpenAPI

  • export:
  • img.png
  • this is API document which client can refer.
  • generate SDK
  • for client to consume API
  • api-gateway-open-api-1.json
  • Replace {region} with your AWS region
  • Replace {lambda-arn} with the ARN of the Lambda function handling the request.
  • check schema. 👈
  • import this into api-gateway
  • schema validation

5. Monitor

CW:logs :

  • enable/disable
  • set log level
  • error only
  • error + info
  • full req + response log

CW:x-rays

  • enable/disable
  • give full picture

CW:metric 👈

  • enable/disable below metrics
  • CacheHitCount
  • CacheMissCount: efficiency of the cache
  • IntegrationLatency (backend latency, latency from lambda)
  • latency (overall : backenf + gateway itself)
  • 4XXError (client-side)
    • 429 throttleError
  • 5XXError (server-side)
    • 504 server timeout
    • 503 service not available
    • 502 bad gateway

B. API-gateway : API key

  • purpose:
  • authentication
  • usage tracking
  • create API keys to identify API clients/customer
    • Ex: WBjHxNtoAb4WPKBC7cGm64CBibIb24b4jt8jJHo9
    • httpheader: x-api-key=your-api-key

C. API-gateway : usage plan

  • create plan at API key level.
  • eg: set:
  • Throttling limits
  • Quotas limits is the overall number of maximum requests
  • associate stages and methods

D. Other 2 API gateway

  • so far working on REST API

D.1. HTTP gateway

  • low-latency | cost-effective just more simple
  • native support for OIDC and OAuth 2.0 + CORS
  • Always proxy
  • HTTP proxy
  • Lambda proxy
  • NO_PROXY ❌
    • hence cannot define any mapping template
  • usage-plans and API-keys ❌

D.2. websocket gateway

  • websocket
  • Two-way interactive communication between a user’s browser and a server : persistent connection
    • Server can push information to the client
    • This enables stateful application use cases

  • use-case : realtime applications

    • chat applications,
    • collaboration platforms,
    • multiplayer games,
    • financial trading platforms.
    • ...

  • WebSocket URL

  • wss://[some-uniqueid].execute-api.[region].amazonaws.com/[stage-name]
  • Communication (connection-id reused)
  • client ==> server
    • img_1.png

  • server ==> client

    • img_2.png
    • operations by server on client:
    • POST : Sends a message from the Server to the connected WS Client 👈
    • GET : Gets the latest connection status of the connected WS Client
    • DELETE : Disconnect the connected Client from the WS connection

  • websocket gateway : backend:

  • lambda
  • DynamoDB
  • Any Http backend