I am new to Spryker and have a General Question: How do we decide when to use Client, Yves, Zed, Ser

U03RJQT5AN4

I am new to Spryker and have a General Question:
How do we decide when to use Client, Yves, Zed, Service, Shared etc. I understand Yves is used for frontend funcitonality of a module and Zed for Backoffice part. But Im not sure if this is correct understanding.
The first module we are making is a common Class which would be responsible for API logging(Request and Response). This class would extend Guzzle Cient and will be used to call APIs instead of guzzle client. We want this class to be used instead of default Guzzle Client to have some control over API calls.(Example logging the calls) Where shall we keep the Class? IMO it should be in a client. How do we decide the best approach and folder.

Alberto Reyer

Shared: Information that will be used in the context of Zed, Yves and Glue, especially transfer definitions and constants

Zed: Backoffice functionality and all business logic

Yves: No logic, only rendering of data that is retrieved from Storage (Redis), Search (Elasticsearch) or Zed (through Gateway Controllers)

Client: Retrieve data from Storage, Search or from Zed (using the Zed Stub)

Service: Wrapped Libraries and general functionalities that are reused in Zed, Yves and Glue

So in your case I would either implement it in Shared or even better in Service to provide a general HTTP Client/API service that is able to log request and responses

So for your case

Chemaclass

As a rule of thumb, this might help you 🧠
• Zed -> Pure backend stuff. It communicates to Postgress (DB) and keep syncronise the data between pgsql and redis/elasticsearch
• Yves -> The door to the FE. It’s the shop what the customer will visit. It fetches data from elasticsearch (searching engine) and redis (key:value db). The only exception that communicates to the posgress db is during checkout proccess. It communicates to external dbs using the Client
• Client -> it’s what Yves uses to communicates to Zed/DB (via Stubs), or redis (via Storage)
• Service -> isolated logic that should be agnostic from Yves or Zed. It could be called from anywhere.
• Shared -> It’s for shared classes from anywhere in the project. Usually constants that you might want to access from your Config (and config_default.php), or the definition of your transfer.xml objects (that later you can find inside the Generated folder)

giovanni.piemontese

@U015S0C0V29

The only exception that communicates to the posgress db is during checkout proccess. It communicates to external dbs using the Client

What means with the only exception ?
Yves has no possibility to communicate with db, also during checkout process.

Yves use always Client to communicate with db.

Chemaclass

Totally 💯. Probably I didn’t express myself properly. I meant, Yves uses the Client to talk to Redis and Elastic for everything. The “only exception” where it communicates to the DB (also via the Client using Stubs) is during the checkout process.

U03RJQT5AN4

Understood. Thanks for clarifying guys. I thought Client is used for communication so it would be better to keep API logging there. But services makes sense too. I would use that.

U03RJQT5AN4

Also, which lib shall we use for generating logs. Spryker docs dont mention any. Upon some RnD, I've found spryker/log(https://packagist.org/packages/spryker/log) and monolog. But im not sure if they are used at application level. Am I missing something?

Alberto Reyer

@UL65CH0MC Just for completeness, there is a possibility that Yves reads directly from the DB (https://docs.spryker.com/docs/scos/dev/tutorials-and-howtos/howtos/howto-replace-key-value-storage-with-database.html), at least as a replacement for the redis storage.
This will technically enable Yves to use the DB connection for other purposes as well.

But it's not recommended and only meant to for cases with a low number of customers, but a high amount of product/availability/price/etc. data.

giovanni.piemontese

@UL6DGRULR Yes I know this exception!!! Good that u mentioned it!