We have an API to fetch the latest transaction data of the user based on the scheduled Next_Refresh_Time. Each user has different scheduled refresh time. Since we have thousands of users we have to run the scheduler to fetch the data. Please suggest me the best way to do it.
Each user has different scheduled refresh time. Since we have thousands of users we have to run the scheduler to fetch the data.
You could add a queue message and specify initialVisibilityDelay with Next_Refresh_Time value when a user login, and then you could create and run a Queue-trigger WebJob to process queue message and featch the latest data (and if current user is still online, add the message (specify same content and initialVisibilityDelay as original message) to queue).
Besides, if you’d like to real-time push the latest data to specific connected user, SignalR would help you implement real-time functionality and SignalR can be used in a variety of client platforms. You could save connection id of a login user in queue message, and then you can call hub method in WebJob function to push data to a connected user based on connection id.
The following thread and article would be helpful to know how to establish connection and call hub method.
SignalR - Broadcasting over a Hub in another Project from outside of
a
Hub
Hubs API for
SignalR
Related
I'm developing an API that calls subscription and tenant information when a user logs in to azure through azure-cli.
When a user logs in to CSP(internally, CSP login is performed using az login --use-device-code), the user receives the url https://microsoft.com/devicelogin and a secret code.
After completing this process, the user's login information is left on our main server, and the user's subscription and tenant information can be called using this.
In the case of a single user, this process has no problem at all. But, when multiple users make a request at the same time, the information of only the user who made the last request is called.
In addition, another problem is that the main server's process is blocked until the process is finished when requested via azure-cli.
When multiple users request CSP login, what is the way to operate normally without blocking the main server and without overwriting user information?
For reference, the main server is FastAPI.
Based on what you have explained, it looks like you have a fast api server which exposes an api which the users use for login. The response from the api will be the device-login url and the device code for the client to use. All the subsequent calls will be done using this information.
So now the first problem is the main thread issue, where the api should not get block until the process is done.
For this you need to have asynchronous apis, which performs background jobs and inform the user when the task is complete or the user can request for the information periodically until the job status changes.
This means we will need to modify our fastapi server with something that can perform multi-threading, queue tasks and have some lookup cache.
The best approach I suggest is having a fastapi stack with the below libs:
Celery - an asynchronous task manager that lets you run and manage jobs in a queue.
RabbitMQ - a message broker that is used to communicate between the task workers and Celery
Redis - an in-memory cache (key-value store) for storing and retrieving values.
You can use the same Redis cache to solve your second problem asw well where you need to store and use multiple session information without overwriting. This will help you handle multiple session info or multiple user infor.
Please have a look at some sample projects from github below:
jjpizarro/fastapi-celery-1
CloudNua/fastapi-celery
Scenario
When business transactions are performed, we're supposed to make that data available to end clients.
Current Design
Our web app publishes transaction messages are added to a topic on the Azure Service Bus.
We expose APIs to clients through which they can consume the data from those transactions.
Upon calling these APIs, we read the messages from the Subscription and return it to the client.
Problem
We want a guaranteed delivery - we want to make sure the client acknowledges the delivery of the data. So we don't want to remove the message from the subscription immediately. We want to keep it until the client acknowledges it.
So we only want to do a "Peek" instead of "Receive".
So the client calls the first API, to get the data, where we do a Peek.
And once the client has received the packets, the client would call a second API, to acknowledge.
At this point, we want to remove the message from the Subscription, making it Complete.
The current design of the Service Bus Message Receiver is that, a Complete can be performed only using the same Receiver instance that performed the Peek, as per the documentation, and we also observed the same when we tried it out.
Both the APIs, are two separate APIs and we cannot do the Peek and Complete using the same instance of the Receiver.
Thinking about options to somehow make the Receiver as a Singleton, across APIs within that App Service.
However this will be a problem when the App Service scales out.
Is there a different way to achieve what we're trying to do here ?
There is an option available in Azure Service Bus to defer messages. Once a message is deferred, it can be received with the help of it's sequence number.
The first client should receive the message and instead of completing it, it should defer it and return it.
The second client (which has sequence number) can receive the message from the Subscription. Refer here for more details.
Another option would be to not use a Service Bus Client on your backend and instead your clients could directly work with Service Bus using its Service REST API (assuming they can't use the AMQP client if I am understanding your scenario correctly).
There are APIs to
Peek-Lock
Renew Lock
Unlock
Delete (Complete)
You could also proxy these requests if you'd like using your backend itself or a service like APIM if you are already using it.
PS: Cross posting the answer for the same query on the MSDN forum
I am creating a bot using Microsoft Bot Framework (BotBuilder) and want it to message the user when an appointment is about to begin.
I currently use Microsoft Graph api to access the user's Office 365 calendar and store the appointments. A background thread then keeps track of time and then messages the user when an appointment is about to start.
The current idea is to use Graph webhooks to notify my bot about new appointments.
My question is, would it be smarter to use an Azure service (such as Scheduler) to keep track of the appointments, and send rest messages to my bot, which will then send a message to the user?
My worry is, that as the amount of users rise, the amount of appointments and time checks will become too large, and that maybe Azure services would be able to handle it better.
This is a perfect fit for Azure Functions with a HTTP Trigger.
https://learn.microsoft.com/en-us/azure/azure-functions/functions-bindings-http-webhook
This article explains how to configure and work with HTTP triggers and bindings in Azure Functions. With these, you can use Azure Functions to build serverless APIs and respond to webhooks.
Azure Functions provides the following bindings:
An HTTP trigger lets you invoke a function with an HTTP request. This can be customized to respond to webhooks.
An HTTP output binding allows you to respond to the request.
I'm trying to piece together the general workflow of giving a user push notifications via the service worker.
I have followed this Google Developers service worker push notifications tutorial and am currently thinking about how I can implement this sort of thing in a small user based web app for experimentation.
In my mind, the general workflow of an web app supporting push notifications is as follows:
Client visits app
Service worker yields a push notification endpoint
Client sends the endpoint to the server
Server associates the endpoint with the current user that the endpoint was generated for
Every time something that your app would say is notification worthy happens, the server grabs the push notification endpoint(s) associated with the user, and hits it to send a push notification to any user devices (possibly with a data payload in Chrome 50+, etc)
Basically I just want to confirm that my general implementation thoughts with this technology are accurate, else get feedback if I am missing something.
You are pretty much bang on, there are some specifics that aren't quite right (but this is largely phrasing and may be done to personally taste).
Client visits app
Register a Service Worker that you want to use for push messaging
Use the service worker registration to subscribe the user to push messaging, at which point the user agent will configure an endpoint + additional values for encrypting payloads (If the the user agent supports it).
Client sends the endpoint to the server
Server store the the endpoint and data for later use (The server can associate the endpoint with the current user if the server if the web app has user accounts).
When ever the server wishes to send a notification to a user(s), it grabs the appropriate endpoints and calls them that will wake up the service worker which can then display a notification.
Payload support in coming in Chrome 50+ and at the time of writing payload is support in Firefox, but there are 3 different versions of encryption used for the payloads in 3 different versions of Firefox, so I'd wait for the payload support story to be ironed out a little before using it / relying on it.
I would like to use Azure Notification Hubs to send push notifications to users of my app running across iOS, Android and Windows Phone.
I have managed to get the basics working but I don't know how to manage the App uninstall story.
On starting, the mobile app will call my Identity Svc to get an Auth Token.
It then calls its Platform Notification service (eg Google Cloud Messaging, APNS) to get a PNS Token.
After persisting the token to local storage it will call a back-end Contact Svc to register the customer's device. This service will create a subscription to the Azure Notification hub for the device.
This is illustrated in the following diagram:
Later on a back-end publishing service will call the Contact Service requesting a push notification for a particular user id. The contact service will lookup the Id allocated to a tag on the notification hub and send a push request.
What options are available to determine when a customer uninstalls the app? Is it just a matter of trapping errors when calling "Send" on the notification hub? I guess this could work if only sending to a single user but my intention is that certain message types are to be published to multiple subscribers. On the initial registration of a device a subscription will be created for a tag of the user id but also for a more general tag such as "New Promotion". The publishing service would later want to issue a "New Promotion" notification to all devices.
Why do you need to know app uninstalls?
Notification Hubs automatically expire registrations for devices that get uninstalled.
Also, I would avoid persisting the PNSHandles in your service at all.
The current guidelines for using hubs are the following:
store the registrationIds of the registrations associated with the device in local storage. This enables you to update tags and channel information with a single update call.
Because mobile connections are not always reliable, it is best to avoid creating a new registration without being able to store the registrationId in local storage. This can result in the device registering multiple times, causing duplicate notifications. You can achieve this by using the Create Registration ID and Create or Update Registration REST APIs.
The first API returns a registrationId without actually creating a registration. When the ID has been securely stored on the device storage, the device can call the Create or Update Registration API.
So I would have your ContactSvc expose two functionalities:
a) create registration id (just call hub to get it)
b) create or update registration given (registrationId, pnsHandle, tags?)
Then your device keeps the regId in its storage, ad after getting the handle from PNS, if as regId is not there creates a new one with endpoint a), then updates the registration with pnsHandle and tags.
Note that in this way your service does not need to persist handles or UUIDs, and does not need to worry about app uninstalls.
If you are keeping track of users, one approach is to periodically (once a month?) check in your hub if its registrations are still there...
You can reach me at #eliodamaggio if this is not clear.