I know it invokes different Lambda instances for different routes (like connect, disconnect, default, etc) on the Websocket API. But what happens for different messages on the same route, does it keep the Lambda instance running for new messages until disconnect?
Let's say, I am building a login form with 2FA. I take username, password and process it, and then I want the 2FA code from client. Can I do this with a single Lambda instance?
As commenter deceze wrote:
You can never assume that a single Lambda instance will process a request.
The point of serverless is that you don not manage the servers. Amazon does. And they can and will start new instances of your Lambda, terminate existing instances etc.
So if you need "cross invocation persistence", you need to solve this in a different way. One common way is to use DynamoDB or depending on the use cases ElastiCache, S3, EFS etc.
Related
I'm writing to you because I'm quite a novice with AWS... I only worked before with EC2 instances for simple tasks...
I am currently looking for an AWS service for reciving data using REST API calls (to external AWS services).
So far I have used EC2 where I deployed my library (python) that made calls and stored data in S3.
What more efficient ways does AWS offer for this? some SaaS?
I know that they are still more details to know in order to choose a good services but I would like to know from where I can start looking.
Many thanks in advance :)
I make API requests using AWS Lambda. Specifically, I leave code that makes requests, writes the response to a file and pushes the response object (file) to AWS S3.
You'll need a relative/absolute path to push the files to wherever you want to ingest. By default lambda servers current working directory is: /var/task but you may want to write your files to /tmp/ instead.
You can automate the ingestion process by setting a CloudWatch rule to trigger your function. Sometimes I chain lambda functions when I need to loop requests with changing parameters instead of packing all requests within a single function,
i.e.
I leave the base request (parameterized) in one function and expose the function through an API Gateway endpoint.
I create a second function to call the base function once for each value I need by using the Event object (which is the JSON body of a regular request). This data will replace parameters within the base function.
I automate the second function.
Tip:
Lambda sometimes will run your requests inside the same server. So if you're continuously running these for testing the server may have files from past calls that you don't want, so I usually have a clean-up step at the beginning of my functions that iterates through my filesystem to make sure there are no files before making the requests.
Using python 3.8 as a runtime I use the requests module to send the request, I write the file and use boto3 to push the response object to an aws S3 bucket.
To invoke an external service you need some "compute resources" to run your client. Under compute resources in aws we understand ec2, ecs (docker container) or lambda (serverless - my favorite)
You had your code already running on EC2 so you should already know you need VPC with a public subnet and ip address to make an outbound call regardless the compute resource you choose
Currently, I have a working modded Minecraft server working running on a C5 EC2 instance. The problem is that I have to manually start and stop the server which can be annoying for my friends. I was wondering if it would be possible to automate the EC2 state so that it runs as soon as a player attempts to join the sever. This would be similar to how Minecraft realms behaves which I heard Mojang was using AWS for:
https://aws.amazon.com/blogs/aws/hosting-minecraft-realms-on-aws/
I have looked up tutorials for this and this is the best I could come across:
https://github.com/trevor-laher/OnDemandMinecraft
The problem with this solution is that it requires to make a separate website to log users in and start the EC2 instance while I want the startup and shutdown to be completely automatic.
I would appreciate any guidance.
If the server is off, it would not be possible to "connect" to the server. Therefore, another mechanism is required that can be used to start the server.
Combine that with your desire to minimise costs and the only real solution is to somehow trigger an AWS Lambda function, which could start the server.
There are a few ways you could have users trigger the AWS Lambda function:
Make a call to API Gateway
Upload an object to Amazon S3
Somehow put a message in an SNS topic or an SQS queue
Trigger an Amazon CloudWatch Alarm (which calls Lambda via SNS)
...and probably other ways
When considering a method to use, you should consider security implications such as:
Whether only authorized users should be able to trigger the Lambda function, or is it okay that anybody (eg a web crawler) might trigger it.
Whether you willing to give your friends AWS credentials (not a good idea) that they could use to start the server directly, or whether it should be an indirect method.
Frankly, I would recommend the following architecture:
Create an AWS Lambda function that turns on the server
Create an API Gateway that triggers the Lambda function
Give a URL to your friends that calls the API Gateway and passes a 'secret' (effectively a password)
The API Gateway will call the Lambda function, passing the secret
The Lambda function confirms that the secret is correct and starts the Amazon EC2 instance with Minecraft installed
Here is a tutorial that shows many of these concepts: Build an API Gateway API with Lambda Integration
The purpose of the secret is to avoid the server from starting if an unauthorized person (or a bot) happens to hit the API Gateway endpoint. They will not provide the secret, so the server will not be started.
Stopping the server after a period of non-use is a different matter. The library you referenced might be able to assist with finding a way to do this. You could have a script running on the Minecraft server that monitors the game and, after a period of inactivity, simply calls the operating system to perform a Shutdown.
You could use a BungeeCord hub server that then allows user to begin a connection to the main server and spin it up via. AWS.
This would require the bungee server to be always up however, but the cost of hosting a small bungee server should be relatively cheap.
I don't think there's any way you could do this without having a small server that receives the initial request to spin up the AWS machine.
I am creating a serverless infrastructure with multiple functions. So far I have managed to publish a new function on AWS lambda using the aws-sam-cli.
One of the last functions is my firebase listener which is supposed to trigger certain aws lambda functions.
Intially, I thought to create a new function and add the listener as follows:
import firebase_admin
cred = firebase_admin.credentials.Certificate(cert_json)
app = firebase_admin.initialize_app(cred, config)
bucket = storage.bucket(app=app)
node_to_listen = '/alerts/'
firebase_admin.db.reference(node_to_listen).listen(listener)
However, the issue is that AWS lambda seems to be designed not to run functions continuously but only be triggered by events. This is true as well for the Firebase listen() function, which means that we get a chicken or egg problem, who triggers who?
How can I therefore publish the firebase listener function and where? Should it be deployed somewhere else (e.g. Heroku?) in order to continuously listen and send the event requests to aws lambda? Or is there a way to connect those two?
There's no way to keep an active listener in any Functions-as-a-Service environment that I know of. The whole purpose of such environments is to run (short) workloads in response to events. You are trying to actually to trigger an event by keeping a listener, which simply doesn't fit the FaaS model.
The two solutions I can see:
Implement your listener on an environment that keeps an active process.
Implement your listener on a FaaS environment that can itself listen to Firebase Realtime Database events. The only environment that can current do so it Cloud Functions, which has Firebase Realtime Database as an event source. So you'd then trigger your Lambda function from Cloud Functions.
The second solution is the only one that really feels fully serverless, but it seems a bit weird to trigger Amazon Lambda from Google Cloud Functions.
There is work under way to allow interop between FaaS providers. But I'm not sure of the current status (link to spec/working group welcome), nor if your scenario would be covered in there.
I'm currently setting up a small Lambda to take snapshots of all the important volumes of our EC2 instances. To guarantee application consistency I need to trigger actions inside the instances: One to quiesce the application before the snapshot and one to wake it up again after the snapshot is done. So far I have no clue how to do this.
I've thought about using SNS or SQS to notify the instances about start and stop of the snapshot, but that has several problems:
I'll need to install (and develop) a custom listener inside the instances.
I'll not get feedback if the quiescing/wake-up is done.
So here's my question: How can I trigger an action inside an instance from an Lambda?
But maybe I'm approaching this from the wrong direction. Is there really no simple backup solution? I know azure has a snapshot based backup service that can do application consitent backups. Did I just miss an equivalent AWS service?
Edit 1:
Ok, it looks like the feature 'Run Command' of AWS Systems Manager is what I really need. It allows me to run scripts, Ansible playbooks and more inside an EC2 instance. When I've got a working solution I'll post the necessary steps.
You can trigger a Lambda function on demand:
Using AWS Lambda with Amazon API Gateway (On-Demand Over HTTPS)
You can invoke AWS Lambda functions over HTTPS. You can do this by
defining a custom REST API and endpoint using Amazon API Gateway, and
then mapping individual methods, such as GET and PUT, to specific
Lambda functions. Alternatively, you could add a special method named
ANY to map all supported methods (GET, POST, PATCH, DELETE) to your
Lambda function. When you send an HTTPS request to the API endpoint,
the Amazon API Gateway service invokes the corresponding Lambda
function. For more information about the ANY method, see Step 3:
Create a Simple Microservice using Lambda and API Gateway.
I'm hoping to use Amazon Lambda to run some background tasks for my web app. These particular tasks will only need to run once for the app (not once per user), so I'd like any user to see in the UI if a task is already running, and I'd like to disable the UI that allows them to start that task again.
Does Lambda offer a way to check the status of a function to see if it is running? If not, what is the best way to persist this info to my web app? Am I taking the wrong approach here altogether?
Lambda functions are supposed to be stateless and keeping functions stateless enables AWS Lambda to rapidly launch as many copies of the function as needed to scale to the rate of incoming events. While AWS Lambda’s programming model is stateless, your code can access stateful data by calling other web services, such as Amazon S3 or Amazon DynamoDB.