Is it possible to update configuration settings programmatically? - azure

We are experimenting with deploying an MVC app as an Azure web role. Currently the app is being hosted locally on our server. It has a few appSettings in Web.Config that can be changed by the users as part of the "Administration" module of the application.
I know this isn't a good practice for Azure because there will potentially be multiple instances of the application running with multiple Web.Configs, which makes updating them all a nightmare (if not impossible). My understanding is that the ConfigurationSettings specified in the service definition should be used instead of Web.Config so that settings are defined globally in one place that all the instances of the application can access.
My question is is it possible to programmatically update ConfigurationSettings similar to the way we update Web.Config settings, or would it be better for us to move those settings into a database or something else?

Yes, but unfortunately it is definitely not simple to do.
Follow this URL: http://msdn.microsoft.com/en-us/library/windowsazure/ee460809.aspx
It talks about a Svc Management API call that you can make to read/write the Service Configuration. It is a 64-base string which you'll need to decode, find XML flags in it that you want to change and re-encode it back and send it back to the API.
Not pleasant, but doable.

Related

What is the best practice for updating an already existing web app deployment using ARM?

My company developed an Azure Resource Manager-based solution that deploys a set of resources (essentially a Storage, SQL DB and Web App), and it is already implemented as our provisioning process for new customers.
However, we are now studying the best way to perform updates, and one of the hypotheses we are considering is having a specific template that updates the binaries of this application.
The idea is to have a separate template, that only has the web app, an app host and a MSDeploy resource that gets the latest version of our package and reuploads it to that web app.
The only problem I'm seeing with this solution is the ability to handle any changes in configuration that are necessary with newer version of the binaries - we do not want users to have to re-input any parameters they placed for the original deploy (done via a Deploy To Azure button), so, any configurations will have to be performed within the application - the plan is for it to use the Microsoft.WindowsAzure.Management.WebSites library.
The major limitation with using Microsoft.WindowsAzure.Management.WebSites is that you are restricted to authenticating with either a certificate or a service principal. Ideally we would like to find a way for the updates to not need any authentication other than the one you provide when you are deploying the update.
Is there any recommendation of best practices to follow for this kind of scenario?
Thank you.
Link to the equivalent discussion on TechNet
It is possible to update only via ARM templates.
For example connection strings can be added automatically to the application settings even when creating the dependent resources themselves.
Ex. Account storage connection string.
Only first time creation of your web sites will take a bit more time, something like 30 sec.
ARM will not destroy your WebApps if they exist already. it will update only.
If there are no changes, then the deployment is very fast.
If your changes require a new Appsettings parameter, you can enter it in ARM , check in to your repository.
and next deployment will pick up and update the WebApp.
So no need for anyone to log-in and update.
Our final decision was to give up on using ARM exclusively. The Service Principal solution, through the SDK, would allow us to use a Web Job or a Site Extension to perform (automatic or prompted) updates that included configuration changes. However, it would require "too many" privileges - why would a customer trust an application that can, at will, create new resources or update existing ones to increase his Azure bill?
The decision was made to utilize Powershell only for updates - if the customer can see the scripts and authenticate himself, this is not a concern. Sadly, this increases update complexity, but we found it to be a necessary evil.

Using Azure MobileServices library with my own LAN WebApi

I am currently doing some research for the development of a mobile application for our company that should support offline data sync (on an iPad). We have explored many possibilities including PhoneGap/Cordova, Xamarin and simply native iOS development. Xamarin, for many different reasons, seems to be our best choice, so my question will assume we will develop in Xamarin.
I was looking into a library for managing offline data synchronization and the most obvious solution is Microsoft Azure MobileServices. However, my company is Canadian, and apparently it's hard to trust (legally) our data to clouds based in the US. Since we already deployed internally our WebApi on our intranet, I figured there was probably a way to point the MobileServices library to our own WebApi. I have read about the Azure Hybrid Connection possibility, but our data still conveying through Microsoft servers might not be a possibility. So, my question is this:
Is there a way to configure the Microsoft.WindowsAzure.MobileServices Client library to point directly to our intranet, RESTful WebApi backend, without going through any Microsoft Azure servers ?
I understand that, in order to be able to use the Client librairies seamlessly, we probably would have to adapt our WebApi to implement the necessary .net Backend interfaces. I'm mostly wondering if it's even possible as the MSDN documentation on the libraries all seem to point to direct connections to their servers (no possibilities to configure your own connection strings) and all instructions redirect you to their Azure Mobile Services website.
Thank you.
If you look at the API for your mobile client, you'll notice that the Azure Mobile Services Client SDK only cares about two things:
new AzureMobileClient( url, appkey)
...where it's hosted shouldn't be a concern. Everything else is just configuration.
If you want to host the Azure Mobile Services Backend on your own servers, technically you could do this, but there are likely a few caveats. Microsoft has announced that they will be launching a Canadian Azure data center, but we won't see it until 2016.
In the meantime, here's how you can host the services locally. Note that I have not tried to emulate all of the features of Azure Mobile Services (aka Zumo) so your mileage (or kilometerage) will vary.
Hosting Locally:
From a technical feasibility, you absolutely can run the services locally. I know this because you can create the Azure Mobile Services Backend project from within Visual Studio and run it locally for development purposes. This is what our development team does for testing their mobile applications.
Note that you can create the Azure Mobile Service backend directly from within Visual Studio: New Project -> Cloud -> Azure Mobile Service. You can also download the exact same template (pre-configured with your URL and ApplicationKey) directly from the Azure dashboard: Create -> Mobile Service.
Obviously, if you're hosting it on your server it will be up to you to configure and use a proper SSL certificate for your site.
ZUMO Permissions:
By default, the security roles on the server are turned off. So if you're locking down any of your methods using the [AuthorizeLevel] attribute these settings will be ignored at runtime. If you need to enable this feature you can do so by modifying the WebApiConfig.Register() method and marking the site as self-hosted: config.SetSelfHosted(true).
Configuration:
From a configuration perspective, the Azure Mobile Service dashboard provides several tabs for configuring Identity, Push Notifications, Connection Strings and App Settings. Sadly, you won't have a dashboard, but all of these settings have a corresponding value in the local web.config. Any value you provide here is automatically overwritten in Azure, but they're used when running locally.
The minimum settings you'll need to configure are listed here. The ApplicationKey you can distribute with your ZuMo client, but the MasterKey is for the Admin authorization level so you'll want to keep that secret. The MobileServiceName is used by the EntityFramework for your database schema and what appears in the URL of your site.
<add key="MS_MobileServiceName" value="myzumosite" />
<add key="MS_MasterKey" value="masterkey" />
<add key="MS_ApplicationKey" value="appkey" />
Values that start with a MS_ prefix map to corresponding values in the Azure Portal. MS_GoogleClientID and MS_GoogleClientSecret map to the Google Identity values in the dashboard, for example.
Any other value in the AppSettings node is immediately accessible via the ApiServices.Settings property and corresponds to the Settings node in the Azure dashboard.
Database connection strings continue to exist in the connectionStrings node. The same is true for azure notification hub.
Database:
Obviously, the database you configure will be up to you as well. Permissions and User accounts are also obvious. There may be some minor differences between the SQL Azure syntax for Entity Framework database migration scripts that you'll need to worry about. (I've discovered the database migration scripts don't work from the Package Manager, but they do work when the database scripts are run when your website starts)
Caveats:
You will not have a nice dashboard for monitoring performance of your site, reviewing logs or changing runtime settings
You will not be able to scale out your site immediately; Scaling and deployment will be your problem
Deployment configuration is your responsibility (Project -> Publish won't be available unless you configure it)
Not sure if you'll be able to use Azure Active Directory as an authentication scheme, though from the sounds of it that won't be a concern. You can write your own authentication providers: Microsoft's Zumo library only supports a handful, but the underlying Owin.Security package that Microsoft uses supports several dozen systems!
Your site will need to be publically visible to your mobile clients
Push Notifications should work, but you will be using Azure's notification hub for this.
I have no idea where ApiServices.Log will go
The easiest path to take would be to:
Create the Mobile Service in Azure to get the notification hub and settings preconfigured
Download the starter site from the dashboard
Configure the web.config as mentioned here.
It's not possible to simply configure WAMS Client library to work with your own WebApi Backend.
But WAMS library is available at github, so I'm sure you can reuse a lot of code from the WAMS project, especially if you want to use a PCL project.
To route your data securly through Azure, you could think about setting up express route. Additionally, for last weeks update, it's possible to apply a custom domain to the WAMS Backend, including your own certificate to secure your connection.

Azure Websites Application Insights - switching config

I have an Azure Website. For the sake of this question, the production version of the website runs on example.com, and a test version of the website runs on sandbox.example.com.
The only difference between the two is that they have different configuration.
At present, they are running under different websites, and I deploy the same website to each azure website via git.
I'd like to separate out the Application Insights data. Is there a technique or process that anyone uses - apart from editing the ApplicationInsights.config file in the sandbox environment post deploy?
Or would using a deployment slot handle this in some way?
There was a new blog post about exactly this today: Application Insights Support for Multiple Environments, Stamps and App Versions.
The destination of the telemetry is determined by the instrumentation
key (iKey), which is sent along with every telemetry message. In the
Application Insights portal, similar events and metrics with the same
iKey are aggregated to give you charts of average durations, event
counts, the sum of users, and so on. The iKey appears in two places in
your project. One is in ApplicationInsights.config:
<InstrumentationKey>94843456-2345-3456-4567-324562759284</InstrumentationKey>
If your application has web pages, the iKey also appears in a script
in the head of every web page. Usually, it’s only coded once in a
master page such as Views\Shared\_Layout.cshtml.
To direct telemetry to different application resources, we can create
several resources with different iKeys. Then we only have to change
the iKeys in the application at each transition in its lifecycle –
along with other configuration data such as connection strings,
certificates, and subscriptions.
The article then goes on how to do this in code, confg, etc:
1) Add iKey as a property in Web.config:
2) Instead of using the iKey from ApplicationInsights.config, we’ll
set it in the code. In global.asax.cs.
To avoid confusion, remove the <InstrumentationKey> node from
ApplicationInsights.config.
3) Configure the web pages to pick up instrumentationKey: "#Microsoft.ApplicationInsights.Extensibility.TelemetryConfiguration.Active.InstrumentationKey". This is
the script usually found in View\Shared\_Layout.cshtml.
4) Don’t forget to update your Web.config with appropriate iKey
configuration during the deployment process. You might devise a way of
setting it appropriately as part of your build, but I’ll leave that to
you.
Found this semi-related question: How to support multiple Azure subscriptions for a single application with application insights this is for using by cloud services, and it works!
Microsoft.ApplicationInsights.Extensibility.TelemetryConfiguration.Active.InstrumentationKey = ConfigurationManager.AppSettings["appInsightsKey"];
I have done this in my unity registertypes method, it works there.
In the Azure portal for websites, on the config tab there is a section called App Settings. You can put your different configuration settings here. When publishing, azure will inject those settings into web.config.
Then just use WebConfigurationManager.AppSettings as you would normally and it will pull the injected values.

Migrating to Azure Web Role - porting the web.config file

I am migrating an existing Web Application to a Web Role. Does this mean that the web.config will be ignored?
I have connection strings, provider details, and tons of other config items in the web.config. Do I have to attempt to port all of this to the Azure ServiceConfiguration.Production.cscfg file?
In development, the Compute Emulator will not be used (since it takes so long), so we will still need the web.config files.
It depends on your use-case.
The cscfg allows you to reconfigure your WebRoles once deployed without creating a new package and redeploying/upgrading/staging+swapping. I especially think of scenarios where continuous integration is used and release procedures are no longer trivial.
If this scenario is not relevant you could just stick with the web.config and build and deploy new packages on configuration changes.
You could also selectively move items you assume that will regularly change in order to benefit from the configurability without having the effort of migrating everything.
I have used wrapper classes which abstract the configuration mechanisms (web.config app settings keys vs cscfg configurationsettings) in order to be able to use the different configuration mechanisms (e.g. check cscfg and fall back to web.config or the other way round or something along that lines).
Update: If you are using a "recent" azure SDK (1.7+) here is CloudConfigurationManager which might do most of the work for you.

Changing/retrieving configuration settings on role start (Azure)

I'm trying to make a service to more easily configure configuration values on Azure applications. Right now, if I want to change a setting that it the same over 7 different environments, I have to change it in 7 different .cscfg files.
My thought is I can create a webservice, that the application will query for its configuration values. The webservice will look in a storage place, like Azure Tables, and return the correct configuration values.
I've been able to integrate this into a deployment script pretty easily (package the app, get the settings, change the cscfg file, deploy). The problem with that is every time you want to change a setting, you have to redeploy.
Finally the question - Is there a way I can retrieve the configuration settings after the application starts, on role start? It would of course need a base set of settings for the app to start. Retrieving the settings from the web service on application start would be good. Any way that I don't have to redeploy the application and that it will retrieve them automatically will work.
Thanks in advance!
Just use the .cscfg for the minimum set (common to all environment) of configurable settings. The use your web services for rest of the configurations. And don't modify your .cscfg. Just have a settings provider that retrieves settings from web service (via polling or message signalling - pub/sub model). And have a reinitialize settings procedure in place for this settings provider and all the services/components that rely on configurable settings.

Resources