How to keep the api alive even if a datasource is not available on loopback 3?

Question

I have a doubt about the data sources: Context: Currently I'm working on a project where my API uses two datasources: A and B .

Sometimes the datasource B has troubles and is not available while A is always available. When B is not available the whole web service collapses.

My question: Is there any way to program the api to keep it working with the part that only evolves the datasource A when the datasource B is not accessible?

Note: I'm working with Loopback 3

Answer 1

lazyConnect:true Will defer connection until you query a model attached to it, and send the client an error without crashing the server if the connection fails.

  "myDatasource": {
    "name": "myDatasource",
    "host": "ds.com",
    "database": "db",
    "username": "root",
    "password": "",
    "connector": "postgres",
    "lazyConnect": true
  },

My question: Is there any way to program the api to keep it working with the part that only evolves the datasource A when the datasource B is not accessible?

You can use your datasource's events to know when to swap models. Here is something I tested briefly.

server/boot/swap.js

function swapModelDatasource(app, model, ds) {
    const name = model.name;
    app.deleteModelByName(name);
    const m = app.model(ds.createModel(name, model.definition.properties, {
        settings: model.settings,
        relations: model.settings.relations,
        acls: model.settings.acls
    }));
}

module.exports = app => {
    const ds1 = app.datasources.aws;
    const m = app.models.Node;
    ds1.on('connected', () => swapModelDatasource(app, m, ds1));
}