Clarifications about edge functions

[antoinebasseto]

Hello!
I am new to web dev in general which makes me a bit confused about some things, so I am looking for help.

I am working on setting up Stripe payments in a Streamlit app, using SQLite cloud as my database. In order to correctly process payments, Stripe says you need to listen to their web hooks.
Because I am using Streamlit, and I am planning on deploying through their Cloud feature, I am not hosting a backend somewhere and can't listen to web hooks directly.
This led me to look into serverless functions, and then realize that SQLite cloud has them (afaik edge functions are just a special kind of Serverless function where the server they reside in is in proximity to the one your db is in right?).

I see three options before me:

1. When a payment is made, Stripe sends a web hook, it triggers a sqlite cloud edge function, the edge function validates the event's signature, it updates the database accordingly and sends a response.

• Can I trigger an edge function with one of Stripe's web hook?
• Importing modules in edge functions isn't supported yet, I'd have to manually verify stripe's signature, which rings alarm bells in my head of bad practice but maybe I am overreacting?

2. When a payment is made, Stripe sends a web hook, it triggers an AWS Event Bridge, validating the event's signature, the AWS Event bridge triggers an sqlite cloud edge function, which updates the database accordingly and sends a response.

• Can I trigger an edge function through AWS Event Bridge?

3. When a payment is made, Stripe sends a web hook, it triggers an AWS Event Bridge, validating the event's signature, the AWS Event bridge triggers an AWS Lambda function, which connects to the sqlite cloud database and updates it accordingly and sends a response.

• How do I connect to my sqlite cloud database through AWS Lambda?

I guess my main question is one of web dev about which of these seem to make the most sense to someone more experienced in building systems like these? My intuition tells me that reducing the number of tools used is good, which makes me opt for option 1, but that manual verification is bad, which makes me confused between options 2 and 3.

Any help would be appreciated!

[alelazcano]

Hi @antoinebasseto

SQLiteCloud cloud functions don't support the use of external modules, only their own methods and act as a kind of "very basic, precarious middleware" between the outside world and your database.

For example, here's a collection of examples: https://github.com/sqlitecloud/sqlitecloud-js/tree/main/examples. If you use chatGTP, you can tell it to review those examples, and it will better program a response.

The use case for cloud functions, along with SQLitecloud webhooks, is to be an entry or exit point for information. For example, the case you're describing:

1. You have a platform that issues a webhook (POST) to a URL that will always be open to listen for that information packet.
2. The SQLitecloud webhook receives it, which is linked to a cloud function script, and parses it, taking only some of the fields and translating the JSON fields, such as PersonalDocumentNumber to DocumentID, which is the name of the header in the SQLite table. Or "ammount" to "totalammount," for example.
3. It will save the data to one or more tables you have.
4. You could also have a trigger, and even another notification, so that SQLitecloud can send it to another system (or have that system retrieve the information) that the payment has been approved.

I suggest you create the flowchart and the actions you want. If you create a good prompt, the AI ​​will help you.