I have an instance of IBM Graph I need to write data and query it from multiple applications. Do I need to bind my apps to my instance or do I need to create a separate instance for each app? Ideally, I'd like to share my data between all of my applications
Yes, you can share your instance between all of your applications. You don't need to bind your app to the instance, but you can if you want. When you instantiate an IBM Graph instance you get back a set of credentials and a url similar to this
{
"credentials": {
"apiURL": "https://graphrestify-alpha.ng.bluemix.net/2208V...348/g",
"username": "a530...dba",
"password": "ab7c...327"
}
}
Your app(s) will be using the apiURL together with the username:password combination to issue queries against the service. For example, one application can send a /vertices request to create a vertex and another application can be doing a /gremlin request (gremlin query) to the same graph at the same time. Both applications can be both producers and consumers of data.
Related
How do I create a system user? The goal is using this user as an administrator.
I can only create normal (client) and server (something like client) users, but I'm unable to create a system user.
I tried the Atlas App Services Admin Rest API but there I can do just the same I can here.
I see that a possible solution is adding custom user data but there isn't a way in the console.
Any solution?
A server user is for backend access to your database. This may be useful if you are looking to perform backend database administration.
When you use Atlas App Services, you are creating users for frontend access to user-appropriate data. If you want to create users on the frontend with a role/privilege of 'Administrator', then you need to implement that logic.
You can create custom data for users, which is a separate document linked to the user with additional fields. A simple implementation would be a custom field 'isAdmin'.
A rule would need to be implemented to only give your frontend users access to restricted data once 'isAdmin' === true. For example:
{
"%%true": {"%%user.custom_data.isAdmin"}
}
For more rule examples: https://www.mongodb.com/docs/atlas/app-services/rules/examples/
I have to send data to a rest API via data factory.
I get batches of records from a database, send them in the body of the REST API call, which provides a response that informs the action performed on each record.
I created a linked service to the base API URL, and this linked service does the authentication to the API.
My question is how I use this linked service along with a web activity in a pipeline?
The web activity requires me to enter a full URL, which feels redundant as the base URL is already in the linked service.
The web activity does let me add multiple linked services but I'm unsure why it allows multiple linked services and how this is supposed to work.
I would appreciate expertise regarding how the web activity works with a linked service.
Thanks!
I have an external REST based API that I need to create a connection to in order to retrieve data on a regularly scheduled basis (for BI purposes). This API is fairly robust, and supports around 60 distinct endpoints. Also, this same API is used to access information across multiple client sub-domains (e.g. client1.apisource.com, client2.apisource.com, client3.apisource.com, etc.). In other words, the API endpoints are the same for each client subdomain.
So what I'm trying to figure out is whether it's possible to create a single ADF that contains a complete set of pipeline actions for each endpoint THAT uses a "dynamic" URL based on the client subdomains? In other words...what I'm trying to see if its possible to create a single ADF that can manage a dynamic list of base URLs.
I tried to parameterize the HTTP and REST connections, but this doesn't appear to Is this possible yet. Any thoughts? Thanks!
Here is an example of a Web Activity to call a REST API using parameters and expressions. The URL can be an expression like:
#concat('https://management.azure.com/subscriptions/',pipeline().parameters.SubscriptionID,'/resourceGroups/',pipeline().parameters.ResourceGroup,'/providers/Microsoft.Sql/servers/',pipeline().parameters.Server,'/databases/',pipeline().parameters.DW,'?api-version=2014-04-01')
I am writing a public rest api, that uses api tokens that are one per application that uses it. I want to give the ability for the client to create users of the application, but not restrict them with what a user should be.
So I have API_USERS that need a token to authenticate to do any requests to the api. Lets say its an api for messages. So the messages from one application will be identified by that token, but how can I let the application that is using the api to be able to create their own users so that the messages in the database, are first uniquely identified for the application and then uniquely identified for the different users of the application if the application has such needs.
My idea is that I can just add another field like MESSAGE_OWNER and let the user handle how to create unique token for his users and maybe use his own database for storing users. Would that be a good solution? Is there another way I am not seeing?
One of the things I enjoy about Parse is the REST API which allows me to send push notifications to "channels" from an external website. I am curious if this, or something similar is available with Windows Azure Mobile Services.
From what I can see, the best way to achieve this result is to create a table which stores push notifications on the Azure Mobile Services. Override the onInsert node.js script to send a push when a record is inserted. The insert is coming through the existing REST API which does allow me to interact with the tablet itself.
What I would like is the ability to provide, for example, my iOS Device Token to the API along with various application access credentials to a REST endpoint and use Mobile Services to send the notification on my behalf, while keeping the information on those clients in my own database hosted in SQL Azure or another location.
Thanks for any insight.
Jason
Today this is posible by creating a "virtual" table in Mobile Services which you only use for the scripting layer. Instead of performing request.execute(), you'd just put in the code that calls push.APNS(or whichever platform you're pushing to) to push out to your client. You can use the mssql module to execute a stored procedure or custom sql on the connected database (which could then talk to a different SQL Database if you are storing the data separate from the DB connected to Mobile Services). As far as getting data to the service, just send over any data you need just like if you were going to insert the data into the table (so they will be fields on the ITEM parameter in the insert scripts) or use the optional PARAMETERS parameter when making a call against your MSTABLE object (and they'll be available on the REQUEST.PARAMETERS object in your scripts).
The API feature now can be used instead of the virtual tables. More detailed info in this question:
How to use the Azure Mobile Service API feature