How should I pass application-specific configuration parameters to a WSGI app? - configuration

CherryPy, uwsgi, gunicorn, and others all have different ways to pass parameters for the container, such as port, number of threads, etc. However, my app has its own domain-specific configuration parameters that the container knows nothing about, naturally. I can think of several ways to get that information in, such as file in a well-known location, environment variables (gasp!), etc.
Is there a standard or preferred way to do it? Something in the wsgi protocol that I haven't been able to find? Some option like "--user"? I haven't been able to find "best practices" advice on this.

Related

Good practices for app configuration storage?

We have a number of loosely coupled apps, some in PHP and some in Python.
It would be beneficial to have some centralized place where they could get both global and app-specific configuration information.
Something like, for Python:
conf=config_server.get_params(url='http://config_server/get/My_app/all', auth=my_auth_data)
and then ideally use parameters as potentially nested attributes, eg. conf.APP.URL, conf.GLOBAL.MAX_SALES
I was considering making my own config server app, but wasn't sure, what would be the pros and cons of such approach vs. eg. storing config in centralized database or any other multiple-site accessible mode.
Also, if I perhaps was missing some readily available tool with good support, which could do this (I had a look at Puppet and Ansible, but they seemed to be very evolved tools for doing so much more than this. I also looked at software recommnedation SE for this, but they have a number of such question unanswered already).
I think it would be a good idea for your configuration mechanism not to be hard-coded to obtain configuration data via a particular technology (such as file, web server or database), but rather be able to obtain configuration data from any of several different technologies. I illustrate this with the following pseudo-code examples:
cfg = getConfig("file.cfg"); # from a file
cfg = getConfig("file#file.cfg"); # also from a file
cfg = getConfig("url#http://config_server/file.cfg"); # from the specified URL
cfg = getConfig("exec#getConfigFromDB.py"); # from stdout of command
The parameter passed to getConfig() might be obtained from, say, a command-line option. The "exec#..." format is a flexible mechanism, but carries the potential danger of somebody specifying a malicious command to execute, for example, "exec#rm -rf /".
This approach means you can experiment with whatever you consider to be an ideal source-of-configuration-data technology and later, if you discover that technology to be inappropriate, it will be trivial to discard it and use a different source-of-configuration-data technology instead. Indeed, the decision for which source-of-configuration-data technology to use might vary from one use case/user to another.
I developed a C++ and Java configuration file parser (sorry, no Python or PHP implementations) called Config4*. If you look at chapters 2 (overview of syntax) and 3 (overview of API) of the Config4* Getting Started Guide, you will notice that it supports the kind of flexible approach I discuss in this answer (the "url#... format is not supported, but "exec#curl -sS ..." provides the same functionality). 99 percent of the time, I end up using configuration files, but I find it comforting to know that my applications can trivially switch to using a different-source-of-configuration-data technology whenever the need might arise.

Openwhisk multitenancy on Openshift

I'm trying to install Openwhisk onto Openshift. I've followed the official guide and it worked.
Now the point is that my environment would be a multitenant ecosystem, so let's pretend having two different users (Ux and Uy) who want to run their containers on my openwhisk environment.
I'd like to have the following projects in my openshift:
Core project, that hosts the Openwhisk's Ingress, Controller, Kafka and CouchDB components (maybe also the Invokers?)
UxPRJ project, that hosts only the containers running actions created by Ux (maybe also the Invokers?)
UyPRJ project, that hosts only the containers running actions created by Uy (maybe also the Invokers?)
The following images better explain what I've in mind:
or also:
Is this possible configuration possible?
Looking around, I wasn't able to find anything like that...
Thank you.
The OpenWhisk loadbalancer which assigns functions to invokers does not segregate users in the way you want, but it is possible to do what you want if you modify the loadbalancer. The way it works now is that there is a list of available invokers which form the allowed set of invokers for a function assignment. You could at that point take into consideration a partitioning based on the user and form the allowed set of invokers differently. There are other ways to realize the partitioning you want as well, but all require modification to the OpenWhisk control plane.

What is "Code over configuration"?

I have seen this terms many times on the google code over configuration or configuration over code. I tried on by searching it on google, but still got nothing. Recently i started work it on gulp again the mystery came code over configuration.
Can you please tell me what is both and what is the difference on them?
Since you tagged this with gulp, I'll give you a popular comparision to another tool (Gruunt) to tell the difference.
Grunt’s tasks are configured in a configuration object inside the
Gruntfile, while Gulp’s are coded using a Node style syntax.
taken from here
So basically with configuration you have to give your tool the information it needs to work like it thinks it has to work.
If you focus on code you tell your tool what steps it has to complete directly.
There's quite a bunch of discussion about which one is better. You'll have to have a read and decide what fits your project best.
Code over configuration (followed by gulp) and the opposite configuration over code (followed by grunt) are approaches/principles in software/program development where both gulp and grunt are using for the same thing as automate tasks. The approach refers to developing programs according to typical programming conventions, versus programmer defined configurations. Both approaches has its own context / purpose and not a question of which one is better.
In gulp each tasks are a java-script function, necessarily no configuration involved up-front (although functions can normally take configuration values) and chaining multiple functions together to create a build script. Gulp use node stream. A stream basically continuously flow of data and can be manipulated asynchronously. However in grunt all the tasks are configured in a configuration object in a file and those are run in sequence.
Reference: https://deliciousbrains.com/grunt-vs-gulp-battle-build-tools/
Because you talked about "code" I'll try and give a different perspective.
While answering a question on figuring out IP address from inside of a docker container Docker container IP address from .net project
there are two codes possible
var ipAddress = HttpContext.Request.HttpContext.Connection.LocalIpAddress;
This will give you the IP address at runtime, but, it won't have control over it.
It can also lead to more code in the future to do something with the IP Address. Like feeding a load balancer or the likes.
I'd prefer a configuration over it.
such as
An environment variable with pre-configured IP addresses for each container service. Such as:
WEB_API_1_IP=192.168.0.10
WEB_API_2_IP=192.168.0.11
.
.
.
NETWORK_SUBNET=192.168.0.0/24
a docker-compose that ties the environment variable to IP address of the container. Such as:
version: '3.3'
services:
web_api:
.
.
.
networks:
public_net:
ipv4_address: ${WEB_API_1_IP}
.
and some .net code that links the two and give access within the code.
Host.CreateDefaultBuilder(args)
.ConfigureAppConfiguration((hostingContext, config) =>
{
config.AddEnvironmentVariables();
})
The code that we wrote is about reading through the configuration.
But it gives way better control. Depending on what environment you are running it you could have different environment files.
The subnet the number of machine they are all configured options rather than tricky code which requires more maintenance and is error prone.

Configure applications using environment variables

12-Factor Apps suggest that you configure your application using environment variables. So far, so good. I can easily imagine that this is a good way to do it if you need to set a connection string, e.g.
But what if you have more complex configuration with lots and lots of values? I for sure do not want to have 50+ environment variables, do I?
How could I solve this, and still be compliant to the idea of 12-Factor Apps?
From a quick read of the configure link you provided, I agree with the author's claim that there is a widespread problem, but I am not convinced that their proposed solution is going to always be best. Like you, I don't relish the idea of having to define dozens of environment variables to configure an application. So here are some alternative ideas.
First, read Chapter 2 of the Config4* Getting Started Guide (disclaimer: I am the main author of that software). In particular, notice that its support for what I call adaptive configuration can go a long way towards addressing the concern that you ask about. Is Config4* the ultimate solution? Possibly not, but I think it is a good step in the right direction.
Second, the chances are that whatever application you are developing/maintaining has already settled on a particular configuration technology, such as XML files or Java property files, and it won't be feasible to migrate to using Config4*. This raises the question: is there anything you can do to avoid having a proliferation of, say, XML-based configuration files when you have multiple environments (such as dev, UAT, staging and production) in which the application will be deployed? I have outlined an approach for dealing with this issue in another StackOverflow article.

The use of config file is it equivalent to use of globals?

I've read many times and agree with avoiding the use of globals to keep code orthogonal. Does the use of the config file to keep read only information that your program uses similar to using Globals?
If you're using config files in place of globals, then yes, they are similar.
Config files should only be used in cases where the end-user (presumably a computer-savvy user, like a developer) needs to declare settings for an application or piece of code, while keeping their hands out of the code itself.
My first reaction would be that it is not the same. I think the problem with globals is the read+write scenario. Config-files are readonly (at least in terms of execution).
In the same way constants are not considered bad programming behaviour. Config-files, at least in the way I use them, are just easy-changable constants.
Well, since a config file and a global variable can both have the effect of propagating changes throughout a system - they are roughly similar.
But... in the case of a configuration file that change is usually going to take place in a single, highly-visible (to the developer) location, and global variables can affect change in very sneaky and hard to track down ways -- so in this way the two concepts are not similar.
Having a configuration file ususally helps with DRY concepts, and it shouldn't hurt the orthogonality of the system, either.
Bonus points for using the $25 word 'orthogonal'. I had to look that one up in Wikipedia to find out the non-Euclidean definition.
Configuration files are really meant to be easily editable by the end user as a way of telling the program how to run.
A more specialized form of configuration files, user preferences, are used to remember things between program executions.
Global is related to a unique instance for an object which will never change, whereas config file is used as container for reference values, for objects within the application that can change.
One "global" object will never change during runtime, the other object is initialized through config file, but can change later on.
Actually, those objects not only can change during the lifetime of the application, they can also monitor the config file in order to realize "hot-change" (modification of their value without stopping/restarting the application), if that config file is modified.
They are absolutely not the same or replacements for eachother. A config file, or object can be used non-globally, ie passed explicitly.
You can of course have a global variable that refers to a config object, and that would be defeating the purpose.