---
title: '17 AWS Lambda'
disqus: hackmd
---

:::info
AWS Certified Developer Associate DVA-C01
:::

17 AWS Lambda
===

<style>
img{
/* border: 2px solid red; */
margin-left: auto;
margin-right: auto;
width: 90%;
display: block;
}
</style>

## Table of Contents

[TOC]

Serverless in AWS
---
- serverless is new paradigm in which devs dont manage servers anymore
- they just deploy code and functions
- initially serverless == faas (func as a service) but now its much more
- serverless was pioneered by aws lambda but now includes anything that's remotely managed
- eg. DBs, msging, storage etc.
- serverless dont mean there's no servers
- means u dont manage/provision/see them
- serverless in aws
- aws lambda
- dynamodb
- cognito
- api gateway
- s3
- sns and sqs
- kinesis data firehose
- aurora serverless
- step funcs
- fargate


AWS Lambda
---
### AWS EC2 vs Lambda
- aws ec2
- virtual servers in cloud
- limited by ram and cpu
- continuously running
- scaling means intervention to add/remove servers
- aws lambda
- virtual funcs - no servers to manage
- just provision code and funcs to run
- limited by time
- short executions up to 15mins
- run on-demand
- only billed when your func is running
- scaling automated

### Benefits
- easy pricing
- pay per request and compute time
- free tier of 1,000,000 aws lambda requests and 400,000 gigabyte seconds of compute time
- integrated with whole aws suite of services
- integrated with many programming languages
- easy monitoring through aws cloudwatch
- easy to get more res per funcs
- up to 3gb of ram per func
- increasing ram of func will also inc cpu and network

### Language Support
- nodejs
- python
- java (java 8 compatible)
- c# (.NET core)
- golang
- c#/powershell
- ruby
- custom runtime api
- community supported eg. rust
- NOTE
- docker is not for aws lambda, it's for ecs/fargate

### Main Service Integrations
- api gateway to create rest api and invoke lambda funcs
- kinesis for data transformations on the fly
- dynamodb to create triggers
- when sth happens in db, lambda func triggered
- s3
- lambda func trigger and file created in s3
- cloudfront will be lambda at edge
- eventbridge react to events in aws with lambda funcs
- cloudwatch logs to stream logs anywhr
- sns to react to notifs and sns topics
- sqs to process msgs from sqs queues
- cognito to react to whenever user logs into db

#### Example - Serverless Thumbnail Creation
- img uploaded in s3
- triggers s3 event notif which is a lambda func
- lambda func code generates thumbnail
- thumbnail either pushed to s3 or metadata inserted into dynamodb

#### Example - Serverless CRON Job
- cron - generate jobs every x mins on ec2 instances
- cloudwatch/eventbridge rule to trigger every 1 hour
- integrated with lambda func to perform your task

#### Example - Lambda Pricing
- [Overall pricing info](https://aws.amazon.com/lambda/pricing/)
- pay per calls
- 1st 1,000,000 requests free
- $0.20 per 1mil requests thereafter
- $0.0000002 per req
- pay per duration in increment of 100ms
- 400,000 gb-seconds of compute time per month free
- equals to 400,000 seconds if func is 1gb ram
- equals 3,200,000 seconds if func is 128mb ram
- after that $1 for 600,000 gb-seconds
- usually very cheap to run aws lambda so it's very popular

### Synchronous Invocations
- sync - cli, sdk, api gateway, app lb
- result returned right away
- err handling must happen in client side
- eg. retries, exponential backoff etc.

- client invoke api gateway which then proxies to lambda func which gives resp to gateway then client

#### Sync Invocations - Services
- user invoked
- elastic lb (alb)
- amazon api gateway
- amazon cloudfront
- lambda@edge
- amazon s3 batch
- service invoked
- amazon cognito
- aws step funcs
- other services
- amazon lex
- amazon alexa
- amazon kinesis data firehose
- NOTE
- for dev associate, only need know alb, api gateway, cloudfront, cognito and step funcs

### Lambda Integration with ALB
- to expose lambda func as http(s) endpt can use alb or api gateway
- funcs can also be either invoked through cli or sdk
- lambda func must be registered in target grp


#### Lambda Request: HTTP to JSON
- elb info
- which elb invokes and what's target grp
- http method and path
- query string params as key val pairs
- headers
- body (base64 encoded)

#### Lambda Response: JSON to HTTP

#### ALB Multi-Header Values
- alb can support multi header vals
- alb setting
- when you enable multi-val headers, http headers and query string params that are sent with multiple values are shown as arrays within lambda event and resp objs


### Lambda@Edge
- have deployed a cdn using cloudfront
- what if want to run global aws lambda alongside ea edge location?
- or how to implement request filtering before rching your app?
- use Lambda@edge to deploy lambda funcs alongside cloudfront cdn
- deployed not in specific region but alongside ea region arnd world with cloudfront cdn
- build more responsive apps
- dont manage servers
- lambda is deployed globally
- customise cdn content
- pay only for what you use
- can use lambda to change cloudfront requests and responses
- after cloudfront receives request from viewer
- viewer request
- before cloudfront forwards the request to origin
- origin request
- after cloudfront receives the response from origin
- origin response
- before cloudfront forwards the response to viewer
- viewer response
- can also generate resps to viewers w/o sending req to origin


#### Example - Global Application

#### Use Cases
- website security and privacy
- dynamic web app at edge
- search engine optimisation (SEO)
- intelligently route across origins and data centers
- bot mitigation at edge
- realtime image transformation
- a/b testing
- user auth and authorisation
- user prioritisation
- user tracking and analytics

### Asynchronous Invocations
- for s3, sns, cloudwatch events etc, events are placed in an __event queue__
- lambda attempts to retry on errors
- 3 tries total
- 1 min wait for 1st then 2 mins wait
- make sure processing is idempotent in case of retries
- if func is retried, will see duplicate log entries in cloudwatch logs
- hence can define dlq (dead letter queue)
- sns or sqs
- for failed processing
- need correct iam perms
- async invocations allow you to speed up processing if dont need to wait for result
- Eg. need 1000 files processed

- s3 create file event into event queue
- lambda func read off queue
- tries to process these events
- if fails, will retry (3 times)
- __idempotency__ - in case of retries, results shld always be the same

#### Services
- simple storage service (s3)
- event notifs
- simple notif service (sns)
- cloudwatch events/eventbridge
- codecommit, codepipeline etc.
- cloudwatch logs
- simple email service
- cloudformation
- aws config
- aws iot and iot events
- many more
- NOTE
- for dev associate, only need know s3, sns and eventbridge

### Integrating with Lambda
#### Cloudwatch Events/EventBridge
- 2 ways
- serverless cron OR rate
- eventbridge rule that triggers func every 1h
- codepipeline eventbridge rule
- eg. to detect everytime codepipeline state changes
- invoke lambda on state change

#### S3 Event Notifications
- `S3:ObjectCreated`, `S3:ObjectRemoved`, `S3:ObjectRestore`, `S3:Replication` etc.
- obj name filtering possible
- eg. *.jpg
- use case
- generate thumbnails of imgs uploaded to s3
- s3 event notifs typically deliver events in seconds but can sometimes take min or longer
- if 2 writes made to single non-versioned obj at same time, possible that only single event notif will be sent
- if want to ensure that event notif is sent for every successful write, can enable versioning on bucket

- s3 send to sns with fanout pattern with sqs
- or send to sqs queue with lambda func
- OR s3 event notif async invoke lambda func
- dlq for any errors

__Event Pattern - Metadata Sync__
- metadata inserted into dyanmodb table or rds db

### Lambda - Event Source Mapping
- applies to
- kinesis data streams
- sqs and sqs fifo queue
- dynamodb streams
- common denominator of all these services is that records needs to be polled from src
- lambda needs to ask the service to get the records and then its returned
- your lambda func is invoked synchronously
- 2 categories
- streams
- queues

- lambda internally creates event src mapping that will poll from kinesis and return results
- when event src mapping has data for lambda to process, will invoke lambda func synchronously with event batch

#### Streams - Kinesis and DynamoDB
- event src mapping creates iterator for ea shard and processes items in order
- can read shard from start with new items from beginning or from timestamp
- processed items arent removed from stream
- consumers can read them
- use case
- low traffic - use batch window to accumulate records before processing
- if high traffic
- can process multiple batches in parallel at shard lvl
- up to 10 batches per shard
- in-order processing for ea batch still guaranteed for ea partition key
- https://aws.amazon.com/blogs/compute/new-aws-lambda-scaling-controls-for-kinesis-and-dynamodb-event-sources/


- error handling
- by default entire batch is reprocessed until func succeeds or items in batch expires
- to ensure in-order processing, processing for affected shard is paused until error resolved
- can config event src mapping to
- discard old events
- restrict num of retries
- split batch on error to work arnd lambda timeout issues
- discarded event go to a __destination__

#### Queues - SQS & SQS FIFO
- event src mapping will poll sqs (long polling)
- specify batch size
- 1-10 msgs
- recommended - set queue visibility timeout to 6x the timeout of your lambda func
- to use dlq,
- setup on the sqs queue not lambda
- dlq for lambda is only for async invocations
- or use lambda destination for failures

- sqs queue polled by lamda event src mapping
- whenever batch returned, lambda func invoked synchronously with event batch

__More Information__
- lambda also supports in-order processing for fifo queues
- scales up to num of active msg grps (grp id)
- for standard queues, items arent necessarily processed in order
- lambda scales up to process a standard queue asap
- when error occurs, batches returned to queue as individual items
- might not be processed in diff grping than orig batch
- occasionally event src mapping might receive same time from queue twice even if no func err occured
- ensure idempotent processing for lambda func
- lambda deletes items from queue after processed successfully
- can config src queue to send items to dlq if cant be processed

#### Event Mapper Scaling
- kinesis data streams and dynamodb streams
- 1 lambda invocation per stream shard
- if use parallelisation, up to 10 batches processed per shard simultaneously
- sqs standard
- lambda adds 60 more instances per min to scale up
- up to 1000 batches of msgs processed simultaneously
- sqs fifo
- msgs with same grp id will be processed in order
- lambda func scales to num of active msg grps

### Destinations
- nov 2019 - can now config to send result to a destination
- hard to see result of async funcs hence can send result to a destination
- async invocations - can define dest for successful and failed events
- destination types
- sqs
- sns
- lambda
- eventbridge bus
- NOTE
- aws recommends use of dest instead of dlq now
- though can use both at the same time
- why dest? cuz its newer and allows for more targets
- dlq just allow u to send failure to sqs/sns but dest send both success and failure into sqs/sns/lambda/eventbridge
- https://docs.aws.amazon.com/lambda/latest/dg/invocation-async.html

- __event source mapping__ - for discarded event batches
- sqs
- sns
- NOTE
- can send events to dlq directly from sqs
- https://docs.aws.amazon.com/lambda/latest/dg/invocation-eventsourcemapping.html

### Lambda Permissions
#### Lambda Execution Role (IAM)
- grants lambda func perms to aws services/res
- sample managed policies for lambda
- `AwsLambdaBasicExecutionRole`
- upload logs to cloudwatch
- `AwsLambdaKinesisExecutionRole`
- read from kinesis
- `AwsLambdaDynamoDBExecutionRole`
- read from dynamodb streams
- `AwsLambdaSQSQueueExecutionRole`
- read from sqs
- `AwsLambdaVPCAccessExecutionRole`
- deploy lambda func in vpc
- `AwsXRayDaemonWriteAccess`
- upload trace data to xray
- when you use event src mapping to invoke func, lambda uses execution role to read event data
- best practice
- one lambda execution role per func

#### Lambda Resource Based Policies
- use res-based policies to give other accs and aws services perms to use lambda res
- similar to s3 bucket policies for s3 bucket
- iam principal can access lambda
- if iam policy attached to principal authorises it
- eg. user access
- or if res-based policy authorises
- eg. service access
- when aws services like s3 calls lambda func, res-based policy gives it access

### Lambda Environment Variables
- env var = key/val pair in string form
- adjust func behaviour w/o updating code
- env vars are avail to code
- lambda service adds own system env vars too
- helps to store secrets
- encrypted by kms
- secrets can be encrypted by lambda service key or your own cmk

### Lambda Logging and Monitoring
- cloudwatch logs
- aws lambda execution logs stored in aws cloudwatch logs
- ensure func has execution role with iam policy that authorises writes to cloudwatch logs
- included in lambda basic execution role
- cloudwatch metrics
- lambda metrics are displayed in aws cloudwatch metrics
- will display
- invocations, durations, concurrent executions
- err count, success rates, throttles
- async delivery failures
- iterator age
- kinesis and dynamodb streams

#### Tracing with X-Ray
- enable in lambda config - active tracing
- runs xray daemon for you
- uses aws xray sdk in code
- ensure func has correct iam execution role
- managed policy is called `AWSXRayDaemonWriteAccess`
- env vars to communicate with xray
- `_X_AMZN_TRACE_ID`
- contains missing tracing header
- `AWS_XRAY_CONTEXT_MISSING`
- by default, `LOG_ERROR`
- `AWS_XRAY_DAEMON_ADDRESS`
- xray daemon `IP_ADDRESS:PORT`

### Lambda in VPC
- lambda by default
- by default lambda func is launched outside your own vpc
- in an aws-owned vpc
- hence cannot access res in your vpc
- eg. rds, elasticache, internal elb etc.

- lambda in vpc
- must define vpc id, subnets and security grps
- lambda will create ENI (elastic network interface) in your subnets
- `AWSLambdaVPCAccessExecutionRole` neede to create ths ENI

- rds sec grp must allow access from lambda sec grp

#### VPC - Internet Access
- lambda func in vpc dont have internet access
- deploying lambda func in public subnet dont give it internet access or public ip
- only have internet if u have a __NAT gateway/instance__
- can use vpc endpts to privately access as services w/o a NAT
- NOTE
- cloudwatch logs works even w/o endpt or nat gateway

- func will go through public subnet with nat device
- nat gateway talks to internet gateway of the vpc
- internet gateway gives access to external api
- dynamodb can be accessed either through internet gateway or use vpc endpts

### Function Configuration
- ram
- from 128mb to 3008mb in 64mb increments
- the more ram you add, more vpu credits u get
- at 1792mb, a func has equivalent of 1 full vcpu
- after 1792mb, you get more than 1 cpu and need to use multi-threading in code to benefit from it
- if app is cpu-bound (computation heavy), increase ram if want to improve perf
- timeout
- default 3 seconds, max 900 seconds (15mins)
- anything above 15mins is not a good case of lambda
- btr for fargate or ecs or ec2

### Execution Context
- execution context is temp runtime env that initialises any external dependencies of your labda code
- great for db conns, http clients, sdk clients etc.
- execution context is maintained for some time in anticipation of another lambda func invocation
- next func invocation can reuse the context to execution time and save time in initialising conn objs
- execution context includes `/tmp` dir
- is a space to write files and is avail across executions

#### Example - Initialising outside Handler
- bad
- db conn established at every func invocation


- good
- db conn established once and reused across invocations
- improve perf


#### /tmp space
- use /tmp when
- lambda func needs to download big file to work
- or if func need disk space to perform operations
- max size 512mb
- dir content remains when execution context frozen, providing transient cache that can be used for multiple invocations
- helpful to checkpoint your work
- for permanent persistence of obj (non-temp), use s3
- cannot use /tmp

### Concurrency and Throttling
- concurrency limit - up to 1000 concurrent executions
- can set __reserved concurrency__ at func lvl
- limits num of concurrent executions
- ea invocation over the concurrency limit will trigger a throttle
- throttle behaviour
- if sync invocation
- return ThrottleError - 429
- if async invocation
- retry automatically and then go to dlq
- if need higher limit (>1000), open a support ticket

#### Concurrency Issue
- if dont reserve concurrency (limit) all func concurrency might be used for 1 app
- other apps get throttled
- concurrency limit applies to all funcs in your acc
- if 1 func goes over limit, all others might be throttled


#### Concurrency and Asynchronous Invocations
- if func dont have enough concurrency avail to process all events, extra requests are throttled
- for throttling errors (429) and system errors (500 series), lambda return the event to queue and attempts to run the func again for up to 6 hours
- lots of retries happen due to throttling
- retry interval increases exponentially from 1 second after 1st attempt to max of 5 mins


#### Cold Starts and Provisioned Concurrency
- cold start
- new instance - code loaded and code outside the handler run (init)
- if init is large (code, dependencies etc.), process can take a lot of time
- 1st req served by new instances has higher latency than rest
- users might be unhapppy
- hence use provisioned concurrency
- concurrency allocated before func invoked (in advance)
- so cold start nvr happens and all invocations have low latency
- app auto scaling can manage concurrency
- eg. for schedule or target utilisation
- ensure enough lambda funcs to be ready and minimise cold start prob
- NOTE
- cold starts in vpc are dramatically reduced in oct and nov 2019
- https://aws.amazon.com/blogs/compute/announcing-improved-vpc-networking-for-aws-lambda-functions/

#### Reserved and Provisioned Concurrency
- https://docs.aws.amazon.com/lambda/latest/dg/configuration-concurrency.html
- look at this in own time as its pretty complicated to describe

### Function Dependencies
- if func depends on external libs eg. aws xray sdk, db clients etc., need to install packages alongside your code and zip tgt
- for nodejs, use `npm` and `node_modules` dir
- for python, use `pip --target` options
- for java, include relevant `.jar` files
- upload zip straight to lambda if less than 50mb else to s3 first
- native libs work
- need to be compiled on amazon linux
- aws sdk comes by default with every lambda func

### Lambda and Cloudformation
- 2 ways to do it

#### Inline
- inline funcs very simple
- use code.zipfile property
- cannot include func dependencies with inline funcs
- so just for simple use cases


#### Through S3
- must store lambda zip in s3
- must refer s3 zip location in cloudformation code
- s3 bucket
- s3 key - full path to zip
- s3objectversion if versioned bucket
- if update code in s3 but dont update s3 bucket, s3 key or s3objectversion, cloudformation wont update your func


### Layers
- custom runtimes
- eg. [c++](https://github.com/awslabs/aws-lambda-cpp)
- eg. [rust](https://github.com/awslabs/aws-lambda-rust-runtime)
- other more common use case is to externalise dependencies to reuse them

- if want to change dependencies will have to keep repackaging app and reupload code
- goal is to externalise your app package
- create layer for your libraries that can be referenced from your code so dont need to repackage everytime for dependencies
- other funcs can also reference your layer

### Lambda Versions
- when work on lambda func, we work on $LATEST
- when we're ready to publish a lambda func, we create a ver
- versions are immutable
- have increasing ver numbers
- get their own ARN (amazon resource name)
- version = code + config
- nothing can be changed - immutable
- immutable = cannot change code or env vars or anything afterwards (fixed)
- ea ver of lambda func can be accessed


#### Lambda Aliases
- __aliases__ are pointers to lambda func versions
- can define a dev, test, prod alias and have them point at diff lambda vers
- aliases are mutable
- enable blue/green deployment by assigning weights to lambda funcs
- enable stable config of event triggers/destinations
- aliases have their own ARNs
- aliases cannot reference aliases

- eg. dev alias always pointing to the latest version
- test alias to test v2 of our app
- prod alias to point to v1 ver that we know is stable

### Lambda and CodeDeploy
- codedeploy can help u automate traffic shift for lambda aliases
- feature is integrated within the SAM framework
- hence this will be elaborated more in the SAM framework
- linear - grow traffic every N minutes until 100%
- Linear10PercentEvery3Minutes
- Linear10PercentEvery10Minutes
- canary - try x percent then 100%
- Canary10Percent5Minutes
- Canary10Percent30Minutes
- allatonce - immediate
- can create pre and post traffic hooks to check health of lambda func
- if anything goes wrong traffic hooks can be failing

- want to shift prod alias from v1 to v2
- slowly from 100% to x%

### Lambda Limits per Region
- execution
- memory allocation
- 128mb to 3008mb
- 64mb increments
- max execution time
- 90 seconds (15mins)
- env vars
- 4kb
- disk capacity in the function container
- in /tmp
- for bigger files compared to env vars
- 512mb
- concurrency executions
- 1000
- can be increased
- deployment
- lambda func deployment size
- compress .zip
- 50mb
- size of uncompressed deployment
- code + dependencies
- 250mb
- can use /tmp dir to load other files at startup
- size of env vars
- 4kb

### Best Practices
- perform heavy-duty work outside of func handler
- connect to db outside of func handler
- initialise aws sdk outside of func handler
- pull in dependencies or datasets outside of func handler
- use env vars for
- db conn strings, s3, etc.
- dont put these values in your code
- passwords, sensitive values etc.
- can be encrypted using KMS
- minimise deployment package size to its runtime necessities
- break down func if func is too big
- rmb aws lambda limits
- use layers when need to reuse libraries
- avoid using recursive code, never have a lambda func call itself

### Console
- use blueprint so dont have to setup from scratch

- choose role that lambda func will use to execute its actions

- prints all 3 keys and return

- can change func code in console

- click on test btn to test the func


- execution role for our func created
- has a res summary indicating access to cloudwatch logs

#### Sync Invocations
- pass func called `hello-world`
- in cli binary format to pass in adj val as text but have it converted in vase64 by cli
- resp written in `response.json` by the func

#### Multi-Value Headers
- this time create func from scratch

- config new alb in ec2 console

- routing point to lambda func

- example of an event obj from lambda func

- if visit dns of alb, will download response
- this is because orig no response with content type pointing to text/html

- change up the response obj so hello world is shown as html


- to enable multi val, go to lb target grp and under attributes enable multi value headers

- alb has a res based policy
- is what allows the alb to invoke lambda func

#### Async Invocations
- to invoke async, cannot do through console have to do through cli

- returns status 202 which is for async invocations
- know func is invoked but we're not waiting for results

- func invoked as seen in logs but we did not wait for results to come back
- hence if func has an error, we also wont know abt it
- hence setup a dlq

- edit your async settings

- can specify num of retries and your dlq

- have to go to iam and give your lambda func role write to sqs

- there's a role specially for this

- cloudwatch has several logs with same request ids which reflects the lambda retries

#### Integrating with CloudWatch
- create a new lambda func

- go to eventbridge and create new rule
- choose schedule


- set target to lambda func

- func invoked every min

#### S3 Event Notifs
- create new bucket with default settings

- add new event in bucket settings

- send to lambda func

- now you'll see s3 as a source of lambda func in your func console

- res based policy allows invocation from amazon s3 service into our func

#### Event Mappers
- add trigger sqs


- event mapper dont have role with enough perms to read from lambda

- now your func have enough perms

- can go to queue actions > add message to send demo msg

- lambda func will have triggered from the event mapper
- lambda will show 0 msgs avail since msgs are deleted after processed

![Uploading file..._e9cjikmyh]()
- disable event mapper if not in use so wont incur costs

- now a demo for kinesis
- batch size
- batch window - gather records tgt?


- starting pos
- whr u want to read from the latest data

- on failure dest
- discard data to sqs queue/topic on failure dest?
- how many retries in case of errors?
- max age of record to process?
- split batch on error? in case batch is too big and lambda func timesout


- concurrent batches per shard
- how many processes u want on ea shard
- 1 means all records in shard processed in order
- if set 10 max, have 10 processors that will process data in shard
- also get ordering by partition key id
- NOTE
- dynamodb will have exact same configs

#### Destinations
- set src, condition and dest type

- sqs role also created to allow sendning msgs into sqs queue

- can see dest in console ui

- 2 dest added
- success and failure

- can view success/failure msgs in sqs queue

- failure msg example
- 2 trigger failure fail

#### Permissions
- role created for lambda funcs

- view res based policy from func console > permissions > scroll to res based policy

- res based policy will be empty for lambda funcs with an event mapper
- as sqs is not polling lambda, its lambda polling sqs

#### Env Vars
- have to first import os package in func to use env vars

- scroll down from func code to see env var window

- can set env var security configs too

- go to same window to edit env var

#### Monitoring
- view cloudwatch metrics in monitoring tab of lambda func

- cloudwatch logs grped by log grps for ea func

- can also enable xray in func console
- enable active tracing

#### VPC
- scroll down func console to view vpc settings

- select subnets and sec grp

- need role to create eli to deploy to vpc

#### Execution Context
- edit basic settings in func console to edit ram configs
- and timeout
- set timeout based on what func is supposed to do
- if u know processing might take awhile, inc timeout
- db conn code shld also be outside of your func call instead of within the func itself

#### Concurrency
- scroll down func console to find concurrency window
- can choose to reserv concurrency
- if set to 0, func is always throttled (used for testing)

- can also set provisioned concurrency

#### Function Dependencies
- nodejs example
- require awsxray sdk
- then create s3 client
- dont need to bundle since its alr bundled with lambda func deployments

- install required sdks
- creates node_modules dir and package-lock.json

- choose zip for code entry type in func console

- cannow navigate node_modules in console too

- enable xray active tracing since code is using xray
- also give required execution roles

- service graph created in xray due to usage in lambda func

#### With Cloudformation
- 3 bucket params impt for lambda func



- res
- iam role
- policies

- this is our lambda func
- specify handler
- specify role too (ref to role defined above)
- for code, specify s3 params (also defined above and referenced)
- runtime and timeout
- enable xray through `TracingConfig`

- enable versioning for s3 bucket with your code

- when creating your cloudformation stack, enter details of your s3 bucket with your code


#### Layers
- click on layers in func console and click add layer

- can get python code from aws docs to copy into the func layer
- code has import numpt and scipy.spatial on top

#### Alias
- qualifiers tab has version and alias

- action > publish new version

- can use version tab to change diff func versions in lambda code

- create alias to point to a specific version
- can use alias to change version instead of using version number

- can give weight for each alias
- eg. 10% of users on that alias will be on a specific version

Quiz
---

###### tags: `AWS Developer Associate` `Notes`