วีดีโอ

Great thank you. Email me from the Books Github or email and I'll send you a special gift.

Great I am glad it helped. You certainly are ahead of the curb if you managed to sit through all the videos. Most people quit tutorials after 10 minutes. Understanding the parts that everybody else quits on is the most valuable thing in any learning course.

Usually this is a permissions issue in IAM from Lambda. Check CloudWatch to see if you are getting any errors when your lambda executes.

Yes but it is not trivial. You would need to store the data in a data repository like S3 or DynamoDB. The easiest way is to use DynamoDB as an IoT data repository created in a separate Rule from IoT Core. Then you can use Lambda to fetch the data from Dynamo via a user SQL statement from a GET request to the Lambda function URL from your web code. If you wanted to pass query parameters from you web code for specific search criteria you would also need API Gateway to interface your Lambda function.

Great thank you, email me for your free gift!

That is usually a browser rendering issue. Try it in a different browser or change the XML header info to something compatible with your browser and refresh your browser. I may need to add a more compatible tag int he near future.

Hi Jaun, the first thing to do is to make sure your IoT payload is actually showing up my using the 'monitor' blade in AWS IoT Core. After that make sure the payload is getting into Lambda by outputting the event object (your IoT payload) from a print or console.log depending on which programming environment you are using. If both of these tests are successful it is very likely you are missing the needed permission to access timestream from Lambda in which case go to IAM and make sure you have correctly given all the needed Lambda permissions as discussed in the video,

@@SteveOnIoT Thank you very much for your answer. I attach a screenshot of the shadow of my iot object: i.postimg.cc/fyXvXtrq/sombra.png The variables change from 00 to 01 without problem. I would like to save all those changes that occur in the three variables (and the date) in the timestream database but I think I am not writing the rule correctly

The first thing you need to do is see what error message you are getting from your browsers terminal. In chrome select Use the keyboard shortcuts On Windows and Linux: Ctrl + Shift + J On Mac: Cmd + Option + J. Sometimes all you need to do is refresh your webpage

Usually that is a permissions issue. So check IAM and see that you have allowed "put" objects to S3.

Thank you

I will try to produce something interesting soon. I had to finish my Azure IoT course first

The traditional way is to write server side code like hosting an MQTT broker and database application in a Linux instance in AWS EC2. However, I show you how to do it using AWS Serverless services with WebSockets later in this lecture series. Propagation varies depending on your distance from AWS region datacenter but I am getting 4/5 of a second latency from the west coast to US-East-1.

I remembered to drink my Ovaltine.

@@SteveOnIoT lol I will do that, watch the relevant videos and test the code :)

My old policy still works when I test it

th-cam.com/video/y_EFv7SUVFs/w-d-xo.html

Hi, DDB is never going to return pure JSON without ETL. This isn't a huge deal, just use a Lambda function to convert any service output you have into JSON. A more costly and complicated approach is to use a AWS GLUE catalog for schema discovery and write a script to translate the data back to S3. What is your intention for your design?

thanks, yes there are a lot of little wired things to work through, fortunately they tend to be the same things over and over

Hi, Timestream databases in general look a little strange compared to traditional databases but they are great for fast and massive IoT data ingest. the challenges with retrieving data compared to traditional databases are still an issue

Hi, Yes you can lose the connection over time, usually once data is being sent the connection stays "hot". WebSockets is still much less robust than HTTP

ESP32 is WiFisecured so it doesn't need to use a sha256 fingerprint. x509 RootCA is totally fungible so you can use this one, it is good for any AWS region. /* root CA found at: www.amazontrust.com/repository/AmazonRootCA1.pem fungible between regions. Open cert, free for public consumption */ static const char rootCA[] PROGMEM = R"EOF( -----BEGIN CERTIFICATE----- MIIDQTCCAimgAwIBAgITBmyfz5m/jAo54vB4ikPmljZbyjANBgkqhkiG9w0BAQsF ADA5MQswCQYDVQQGEwJVUzEPMA0GA1UEChMGQW1hem9uMRkwFwYDVQQDExBBbWF6 b24gUm9vdCBDQSAxMB4XDTE1MDUyNjAwMDAwMFoXDTM4MDExNzAwMDAwMFowOTEL MAkGA1UEBhMCVVMxDzANBgNVBAoTBkFtYXpvbjEZMBcGA1UEAxMQQW1hem9uIFJv b3QgQ0EgMTCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBALJ4gHHKeNXj ca9HgFB0fW7Y14h29Jlo91ghYPl0hAEvrAIthtOgQ3pOsqTQNroBvo3bSMgHFzZM 9O6II8c+6zf1tRn4SWiw3te5djgdYZ6k/oI2peVKVuRF4fn9tBb6dNqcmzU5L/qw IFAGbHrQgLKm+a/sRxmPUDgH3KKHOVj4utWp+UhnMJbulHheb4mjUcAwhmahRWa6 VOujw5H5SNz/0egwLX0tdHA114gk957EWW67c4cX8jJGKLhD+rcdqsq08p8kDi1L 93FcXmn/6pUCyziKrlA4b9v7LWIbxcceVOF34GfID5yHI9Y/QCB/IIDEgEw+OyQm jgSubJrIqg0CAwEAAaNCMEAwDwYDVR0TAQH/BAUwAwEB/zAOBgNVHQ8BAf8EBAMC AYYwHQYDVR0OBBYEFIQYzIU07LwMlJQuCFmcx7IQTgoIMA0GCSqGSIb3DQEBCwUA A4IBAQCY8jdaQZChGsV2USggNiMOruYou6r4lK5IpDB/G/wkjUu0yKGX9rbxenDI U5PMCCjjmCXPI6T53iHTfIUJrU6adTrCC2qJeHZERxhlbI1Bjjt/msv0tadQ1wUs N+gDS63pYaACbvXy8MWy7Vu33PqUXHeeE6V/Uq2V8viTO96LXFvKWlJbYK8U90vv o/ufQJVtMVT8QtPHRh8jrdkPSHCa2XV4cdFyQzR1bldZwgJcJmApzyMZFo6IQ6XU 5MsI+yMRQ+hDKXJioaldXgjUkK642M4UwtBV8ob2xJNDd2ZhwLnoQdeXeGADbkpy rqXRfboQnoZsG4q5WTP468SQvvG5 -----END CERTIFICATE----- )EOF";

I fixed the link, try again, thank you

The Host name will be your IoT endpoint if using AWS IoT Core. If using API Gateway your host name will be your API Gateway generated URL. You can email your program to at borsay@gmail.com and I can look at it.

You can send IoT Data to AWS IoT Core with different topic identifiers or have the user identify themselves in the JSON IoT Payload as a key value pair. To distribute those payloads you would need to check that the topic or payload match the key-value pai either the Rules query statement or in a Lambda function 'rule action' linked to IoT Core. From there the data can easily be dispatched to a folder or any other service you like.

Yes, First make sure your S3 bucket is receiving the sensor readings. After this you can debug the issue by opening the developers panel in your browser when you open the index.html page and start sending data, it will report what the problem is with your code. To open the developer panel press (CNTL+Shift+J) in Chrome and (CNTL + Shift + I ) in most other browsers

Hi Martin, I actually do automate the ConnectionID process later in these lectures so that may largely solve your problem. But to address your specific issue it is likely just a function of the serverless architecture that is abstracted that is confusing you. So if you are using Socket.io on NodeJS you are likely going to use AWS EC2 as a "serverful" instance with NodeJS and thus have storage allocated for your EC2 instance. this simplifies things to some degree. This architecture would handle some of these underlying issues for you but generally cost more on AWS EC2 than serverless with S3 and Lambda. If you go serverless then Lambda is going to serve as your backend and you will have to manage the connectionID manually and socket.io is built on websockets as you know. I believe socketIo calls it sessionID instead of connectionId. var socketConnection = io.connect(); socketConnection.on('connect', function() { const sessionID = socketConnection.socket.sessionid; // ... }); As you already have seen, most AWS WebSockets videos use DDB as the connectionID storage destination. This makes sense if you have an unknown number of chat clients to connect to. With IoT prototyping we can get away with simplifying to using the AWS Parameter store for ConnectionID's because we have a known number of clients ahead of time, and thus can pre-allocate resources.

Also you can simply Publish a message to your clients on an alternate topic if you simply need publish messages to clients via MQTT. Both Boto3 and the AWS-SDK on lambda allow MQTT publishing from lambda.

As far as just sending data from AWS IoT to your website simply search my name on the dev.to website. I have written extensive articles on how to do this via a simple GET as well as through Websockets. As related to sending data from a website to your ESP32 device that is basically the same thing in reverse. You would have to construct a lambda that publishes the incoming event object to a topic that your ESP32 device is subscribed to. Then you would use API Gateway to generate an endpoint that you would embed on your website and use to transmit user input from your website to your lambda function. I'm going to add this example to my serverless course on Udemy soon as I have had several requests for this method.

@@SteveOnIoT Thanks for the reply! It would do wonders if you could add it to the course. I have added the course to my Udemy cart; will buy it as soon as you publish the said content. Thank you so much in advance!

@@SteveOnIoT So sorry to bug you sir. Any updates on this? I couldn't wait so I bought your serverless course on udemy. Now I'm waiting 😊

You can email me from the course, I have time to work on your IoT project today

@@SteveOnIoT Thank you so much! I posted a message regarding this on the QnA as I couldn't see your email there. 🙏

Hi, The https WebSocket 'internal' endpoint in the messaging lambda is needed to make the external connection through API Gateway to the wss endpoint on the external website, this is where IoT data is transmitted to the web host. I originally had an unused wss endpoint listed in my connection lambda but this is unimplemented and unnecessary.

@@SteveOnIoT so that's my doubt , about when i have to use de wss endpoint, and when i have to use connectrion url endpoint?,

Yes you have to insert YOUR own wss endpoint in on the website. wss is used for external connections, https is used for internal connections. that's how AWSdeveloped it in API Gateway

wss is for external connection like a website, http for is internal connections like lambda

@@SteveOnIoT so when i want to send msgs to web host, or active listening conection, use the http url to send payload from lambda to APIGW, and the external wss active connections will be listening? the https url?

Easier definitely not, more "effective" maybe. First Cognito and Amplify are unrelated. Cognito is normally used with CloudFront to provide proper security against unauthorized users. As I said in the video, I'm skipping security in order to keep the videos simple. You can watch one of over a thousand AWS and user videos discussing how to add security to a static website in S3 and simply apply those techniques to my example. An easier and free way of decent security would be altering the bucket policy (limit IP range or source) or using file tags, that would be easier compared to Cognito if you are not bothering to use CloudFront. As far as Amplify, that would make a more complex website. The reason to use Amplify is mostly to take advantage of GraphQL. That is something I might cover for asynchronous IoT but normally it is a sperate, and orthogonal to WebSockets. In most instances GrapghQL is going to be both slower and more expensive than WebSockets for IoT on AWS.

@@SteveOnIoT can we make a websocket api with amplify? , ive heard only graphaQL and REST API with amplify

it may be possible but Amplify is specifically made to integrate with GraphQL which is also a bidirectional asynchronous protocol, so it would likely be redundant to use with websockets, and I doubt you would find an implementation using both GraphQL and websockets together. I could be wrong as I haven't used Amplify much as it seems to have a ton of overhead and is almost aways is used by front-end developers and not IoT engineers.

Are you talking about batching IoT data?

@@SteveOnIoT For example, in your video sensor_5's data has been injected to the DB in 2 separate rows instead of one with both values (humidity and temperature) in the same row

@@panagiotisgiannelos3767 You can certainly transform them in Lambda but as you know multi-measure records were announced at re:invent in 2021 but I don't see them implemented. I suspect the Timestream team is fairly sparse as very little new features or material has been released since Timestream came out, and Timestream was already a very delayed release from announcement. If Timestream wanted to release this feature it would be very easy to implement, a simple button option to allow all variables with same incoming timestamp to reside in the same row.

Glad it was helpful!