Django community: RSS
This page, updated regularly, aggregates Community blog posts from the Django community.
-
Djangocon keynote: devops, it is much more than just automation - Kevin van Wilder
(One of the summaries of a talk at the 2014 djangocon.eu.) Kevin van Wilder talks about devops. (Note: I changed the title from "your product is more than the application" to "devops, it is much more than just automation" as that covers the main message much better). Nothing good happens after 14:00 on friday. So don't do a friday deploy. He remembered a quick fix they needed to deploy one friday evening. It actually went OK and it even worked during the press conference demo. 10 days later the site died, though. Who's problem is it? "It worked in development". So, "it is an operations problem now?". Is devops the solution? You cannot really be a "devops person". It is more a movement, a philosophy: lets work together as dev and ops. Partially it is just a name for something that people have been doing for a long time. A useful acronym is CLAMS: Culture. How can we influence co-workers? You cannot directly change culture. But you can change behavior. And behaviour becomes culture. If you do it right, you'll start influencing others to do right, too. We've been influenced ourselves at this conference, right? Haven't we been shown The … -
Djangocon: typography and the web, the state of the art - Idan Gazit
(One of the summaries of a talk at the 2014 djangocon.eu.) Idan Gazit is someone I really like to hear at django conferences. I'll simply point at two earlier talks. Design for developers (2010) and compass and less (2011). There's a nice book "thinking fast and slow". A general principle is that anything you can to to reduce cognitive strain is worthwhile. One of them is making your text readable. This is mentioned explicitly. The printing press was invented 600 years ago and we did a lot of writing beforehand, so we have a lot of experience. There are two kinds of factors: Micro. Typeface, kerning. Macro: measure ("width"), leading ("spacing"), flow of type on a page. There's a lot of common wisdom for books. How to lay it out, the ideal size of margins. But on the web we have a huge problem: we do not have control. There's no fixed page size, we have many screen sizes and screen resolutions and screen quality. But still... there's a lot that we can learn from the existing body of knowledge. Type size. 16 pixels is the default size in browsers. In books it is often smaller. But it is 16 … -
Djangocon: really, really fast django - Christophe Pettus
(One of the summaries of a talk at the 2014 djangocon.eu.) Christophe Pettus works for http://pgexperts.com . So... postgresql expert! I've got two other talks of him postgresql when it is not your job and advanced postgresql in django. How fast django anyway? You hear things like "the orm is slow" or "the template engine is old". So... when in doubt, measure. How high-overhead are django's components? You'll need to have a good test. Just returning an empty response is nonsense. He did a bunch of tests and django holds up pretty well. The overhead is very low. Most ORM operations are O(N) on number of fields. So on average, a 1-field model will be 10 times faster than one with 10 fields. Updating an update? Use the ORM update method instead of looping over items. Much faster. Using manual SQL can be a little bit faster than the ORM, but it is neglegible. Django's basic request loop is plenty fast. Request/response cycles to the database generally swamp everything else. Always do bulk and batch operations instead of iterating. Don't use components you don't need. Often you load a lot of libraries. Do you really need a full REST library … -
Djangocon: from icontains to search - Honza Kral
(One of the summaries of a talk at the 2014 djangocon.eu.) Honza Kral talks about searching, which is a big topic. He'll focus on unstructured search: fulltext search. Search is an interface to your data. Textual search: $ grep -ri django * or: >>> Documents.objects.filter( ... Q(title__icontains='django') | ... Q(body__icontains='django')) Don't do this. None of them are going to scale. And you're going to miss things. If you search for "running" you'll miss "runs", for instance. The first index in world history was the 1230 bible concordance. Just a simple list of biblical words with pointers to the passages in which the words occurs. We can do the same. Technically it is an inverted index. We can do more than just looking up individual words. We can also look at phrases. In those cases you look both at the files the words occur in, but also at their place in those documents. We can do more things when storing words in our index: Leave out common words like 'a' and 'the'. Lowercase everything. Normalize words (normalize various verb variants to just one, normalize single/plural). Look for synonyms. (Fast is almost the same as rapid, for instance). All this happens at … -
Djangocon: taming complexity with django - Ustun Ozgur
(One of the summaries of a talk at the 2014 djangocon.eu.) Ustun Ozgur tells about someone's talk about simplicity, simpleness and complexity. What does "complexity" tell us as django developers? He quotes Richard Gabriel who talks about "habitability" as a characteristic of source code that enables developers that see a piece of code to understand it and to work on it, even if they haven't seen the code before. This is so for django. Complexity should be fought head-on. Look at domain driven design. Keep your architecture close to the actual domain. That helps. Django has a couple of layers of defense against complexity. URLs. Django routes various parts of the url to various applications. The division into seperate applications helps a lot to fight complexity. Models and managers. Django uses several well-known design patters that help structure your code in the right way. But... watch out for "fat" classes that do too much. You could look at mixins and compositions. Or you can move some of the fat class's methods to separate functions. Sometimes no code is the best code. You could put state machine state in a variable (a dict or so) and use that. Data is easier … -
High Performance Django Kickstarter Project
I'm really excited to (finally) announce that we are writing a book! We've been working with Django professionally for a long time (over 7 years now). During that time, we've learned a lot about how to use the framework to build fast and scalable websites. We're bundling up all that knowledge into an e-book called High Performance Django and it is up on Kickstarter today. We're already well into the writing process, so we can tell you a little about what you'll get. The book is heavy on Django tips, but also reaches far beyond Python, showing you how to architect and tune the rest of your stack. It gives you an opinionated battle-tested blueprint utilizing many of the same techniques as high-profile Django sites like Disqus, Instagram, and Pintrest. We hope to get the first edition out in July, but we need your help to make that happen. Writing and editing the book is going to be a massive undertaking and you can help support our effort. Give us your vote of confidence by backing the project today! -
High Performance Django Kickstarter Project
I'm really excited to (finally) announce that we are writing a book! We've been working with Django professionally for a long time (over 7 years now). During that time, we've learned a lot about how to use the framework to build fast and scalable websites. We're bundling up all that knowledge into an e-book called High Performance Django and it is up on Kickstarter today. We're already well into the writing process, so we can tell you a little about what you'll get. The book is heavy on Django tips, but also reaches far beyond Python, showing you how to architect and tune the rest of your stack. It gives you an opinionated battle-tested blueprint utilizing many of the same techniques as high-profile Django sites like Disqus, Instagram, and Pintrest. We hope to get the first edition out in July, but we need your help to make that happen. Writing and editing the book is going to be a massive undertaking and you can help support our effort. Give us your vote of confidence by backing the project today! -
A break from the tutorial: Intuitive stuff
Hey guys! Let's take a break from the bombarding of information, and let's do something interesting. Let's talk about some of the features that are taken for granted in all the giant websites. You may already have noticed them. If not, double back and check them out. We'll also see how to implement them in your website.The "Keep me logged in":Many websites like Facebook and Gmail offer this option at the time of login. What it does is even if you close the browser/tab/window, the next time you open it it opens on your homepage. Even more fascinating, if you try to log in from some other device it you're not logged in, and you even get an alert that someone tried to access your account from an unknown location. Let's see how this is done in steps:When you log in, cookies with your username are stored in your browser. If you close it and reopen it, and access that site again, it checks for that-site-specific cookies. If they are there, they take you to your homepage. Else, you're redirected to the login page.To prevent multiple people accessing the account at the same type, the session is logged in to … -
Djangocon: Django class based views - Leonardo Giordani
(One of the summaries of a talk at 2014 djangocon.eu) Leonardo Giordani gives a survival guide for novices on django's class based views: what they are and how they work. What are they? Django views based on python classes. For many this translates into some django stuff based on some python magic. If you followed the tutorial but don't exactly understand what a view is, class based views can be quite hard to understand. A view is the part of django that gets the request and returns a response. Views used to be monolythic functions. A request comes in and a response goes out. It is hard to replace one specific part of it. If you could split it up, you could replace the individual parts. Object oriented programming (so: classes) is a good way to split up something that's monolythic into several separate methods on an object. That's what django's class based views do. In the case of class based views, it is worth your time to actually read django's source code. Get familiar with the details, it helps you understand what happens. First the as_view() method you see in the urls.py: url(... SomeView.as_view()). It makes sure your view … -
Djangocon: the future of postgresql in django - Marc Tamlyn
(One of the summaries of a talk at 2014 djangocon.eu) Marc Tamlyn is well-known for his succesful kickstarter campaign to improve postgresql support in django. He asked for a show of hands: some 90% of the attendees use postgresql. It is the most popular database for django projects. Most of the core team favours it. It has a wide feature set and it is a proper active (non-oracle-owned) open source project. Wide feature set... but not all of the features are supported by django. You can use several add-ons, but they're not that nicely integrated. So he proposed a kickstarter project for creating django.contrib.postgres. 14k (in UK pounds) was raised by the kickstarter campaign! Wow. He thanked everyone that contributed (a couple of big contributions by companies by name). The core of the project is to add support for a couple of data types, like array, json and hstore. Array can list just about any type. Nested structure is OK. No foreign keys, though. This one is almost ready for django 1.8. HStore is a key-value store. The most requested feature! It is quite similar to django-hstore, which is already a very good add-on. They'll help him with getting HStore … -
Djangocon: web components in django - Xavier Dutreilh
(One of the summaries of a talk at 2014 djangocon.eu) Xavier Dutreilh talks about web development. You probably use libraries like jquery and backbone. And css frameworks like bootrap. And tools like bower, grunt, yeoman and jshint. One of the things you should care about is performance. Especially for mobile devices, you don't want to send over too much css/js. You don't want too exotic and hard technologies, as you as a web developer probably will lack the specialized knowledge. Accessibility is important. Cross-browser functionality, too. But... we often disregard the wishes we enumerated above and simply look for ready-to-use framework and pile them on top of each other. And we hope everything works as we hope and expect. Which it doesn't and we spend lots of time debugging. We inherit technical dept by piling everything up like this. What can we do? Well, we first have to stop making such a mess in the first place and have to start looking at the requirements we figured out. The best tip is to think about individual elements. We could also try to do it lean and mean: custom code. Start by looking at a web page. What does it consist … -
Djangocon: pair up django and web mapping - Mathieu Leplatre
(One of the summaries of a talk at 2014 djangocon.eu) Mathieu Leplatre (who works on makina corpus) says web mapping should be simple and google maps should become unusual. Fundamentals of cartography Projections and postgis. Cartography is based on location: longitude (x) and latitude (y). Normally you use decimal degrees: -180 to +180 longitude, -90 to +90 latitude. Well known from GPS tools. Problem: the earth isn't perfectly round. GPS uses the "WGS 84" reference ellipsoid to standardize the latitude and longitude. Problem 2: you often have to show it on a flat map or flat display. So you have to project the 3D data on a 2D surface. Naturally this means compromises. Projections are for instance the mercator projection that google uses. For more background, examples and especially illustrations, look at mapschool.io Transformation often happens between "WGS 84" (srid=4326, that is the most-used one for storing it in databases) and google mercator for display (srid=3857). Those scary-looking "srid" numbers? Spatial reference ID. But don't worry, those two numbers are almost always the default value for databases and tools. Another fundamental concept is the data format. Basically you have either vector data (points, lines, polygons) or rasters (bitmap data). Vector … -
Djangocon: two short talks about django-oscar and about performance gains
(One of the summaries of a talk at 2014 djangocon.eu) An introduction to django-oscar - David Winterbottom David Winterbottom is, which I just now noticed, the originator of the very useful www.commandlinefu.com website that I've had in in my RSS reader for the past few years. Recommended: you get a useful tip out of it every month or so. More related to django: he's the author of the e-commerce framework 'django-oscar'. In e-commerce, especially B2B (business to business), you get loads and loads of exceptions. He originally worked on a PHP system that basically broke down under all those exceptions to the standard rules. He now works on django-oscar, which is used by a number of very large customers. Ecommerce: products, baskets, orders. Overridable apps. It could be added to your site without modifying anything: it won't take over your whole site. It doesn't use the django admin though, but a custom edit site. Philosophically, oscar isn't an out-of-the box product. It is a generic product, so it cannot know how your tax system works. Which payment provider you use, How you've handled shipping. So you'll have to add that yourself. Oscar provides a base platform you can build upon. … -
Djangocon: distributed systems - Raphael Barrois
(One of the summaries of a talk at 2014 djangocon.eu) Raphael Barrois has this definition for distributed systems: they are a set of autonomous, interconnected services. Possibly a different codebase. Each service with its own database. Why? When? For elaborate systems you can make a nice tidy elaborate architecture. But after two years of bugfixing and feature-adding and changes it inevitably becomes a huge sprawling mess. So... split it up! Into separate components. When would you start doing something distributed-like? When a new feature has to be added that's not an incremental improvement. A feature that is strongly disconnected from the core features. Another reason can be that you install a new version of your main project somewhere, possibly in another datasenter or with a separate database. It still has to be operated by the same ops team, though. Or there could be technical reasons like scaling/sharding or geographic expansion or debundling/upgrading. Building it Where should you start? A good point is to extract generic code. misc, util, tools. Everything that's not business-specific. Perhaps also some basic business components: common models, UI, etc. Put this kind of code into separate models and adapt your deployment process for fast-moving internal dependencies … -
Djangocon: two talks, ecology and healthchecks
(One of the summaries of a talk at 2014 djangocon.eu. Two (short) talks actually.) Django powered ecological data - Jakub Witold Bubnicki Jakub Witold Bubnicki works on data-intensive science. For instance for biological sciences, more and more data is becoming available. Globally. In data intensive research, you often have to integrate and adapt multiple sources. Smaller teams (let anone individual PhD students) can often not create the infrastructure needed. He collects photos from his "camera traps" for two years now. He has to integrate this with geographical data and, for instance, remote temperature datasets. Often, the data includes both a space and a time component (location and date). The collected data needs to be easily accessible, discoverable, shareable and reusable. This is often not the case. So he started designing a django architecture to fix this. He also uses lots of existing applications like metacat (catalogs), geoserver (geographical data) and rasdaman (rasters). Django provides the authentication, filtering, browsing, searching, permissions. It connects to the external applications. The django site itself also has an API so that others can connect to it. Note: they're still working on it, it is not finished. But the generic data browsing and a bunch of … -
Djangocon: visibility for web developers - Bruno Renie
(One of the summaries of a talk at 2014 djangocon.eu) Note beforehand: Bruno Renie had several useful things to say at previous djangocons. I'd like to draw specific attention to his lightning talk last year about settings. He later wrote a more elaborate version on his own blog. Something I'm gonna take a deeper look at in the coming weeks as I'll have to fix something in many of our internal sites :-) He works in a diverse team and in a quite complex infrastructure, so they need to make their infrastructure visible: errors, events, metrics. Errors: easy: use sentry. Events: basically, a log call. The errors and sentry mentioned above are great for developers, but you often need info on the requests that don't fail, too. For this, you need centralized logging: searching on several machines in logfiles doesn't cut it. As an aggregator, they use logstash + eleasticsearch. They use kibana as a frontend. rsyslog/syslog-ng can work with logstash-forwarder ("lumberjack") to forward everything to logstash. In python, you can use a sysloghandler to send everything to the syslog. You have to use structured logging. A formatted string is nice for the logfile, but not if you want to … -
Djangocon keynote: where the wild things are - Aymeric Augustin
(One of the summaries of a talk at 2014 djangocon.eu) Aymeric Augustin talks about app loading. App loading is a step in django's initialization sequence: import all models and populate a cache. It is also a project that improves this step. The trac ticket for this feature is seven years old... Many many people have worked on it and provided patches and branches. The final implementation looks like this: from django.apps import AppConfig class MyConfig(AppConfig): name = ... label = ... verbose_name = ... # and in the settings file: INSTALLED_APPS = ( 'some_app', # Old style. 'yourapp.apps.MyConfig', # Points to the config class. ... All nice, but in practice there were crashes when using the implementation in production with gunicorn. The root cause was that django doesn't have a proper signal that says that it is fully configured and ready. So he started digging into the code. And found out that actually reading all the models and setting everything up occurs at various points in time, depending on how you call it. "runserver" is different from "shell" which is different from "running it from wsgi". At the core is the AppCache that django uses behind the scenes. This looks … -
Djangocon: tuesday lightning talks
(One of the summaries of a talk at 2014 djangocon.eu) Migrate an existing web application to django - Samuel Goldszmidt What to do when you have a PHP/mysql app and want to turn it into a django site? Set up a django site with the latest version (1.7) with migrations included. Add your database settings. Create an app and use manage.py inspectdb > yourapp/models.py to inspect your database and generate models. Then create the default core django tables (just run manage.py migrate). Start the initial migration manage.py makemigrations yourapp. Start renaming tables and fixing things up and turn those steps into migrations. At the end you have a nice clean set of models including automatic migrations. You could use the "squashmigrations" command to get a new clean baseline migration after you're ready. Harness the speed of the wheel - Xavier Fernandez Wheel is python's new (binary) package format. It is a zip format archive, you can basically extract it and it is ready. It has a specially formatted file name that includes information for the target architecture (as it can contained compiled code). Installing from wheels is 4 or 5 times faster for Django. For something that is heavy in … -
Internet & mobile connectivity
Internet & mobile connectivity The conference and sprint venues will of course be furnished with a suitable wireless network. You'll also be able to get wireless access from your hotel apartments and rooms. You may all the same wish to buy a SIM card for mobile connectivity. You need to do this before you arrive on the island. We recommend Orange, Bouygues or SFR. -
Django sticky queryset filters
In Django, Stuff.objects.filter(a=1).filter(b=1) is almost always the same as Stuff.objects.filter(a=1, b=1). Everyone knows and expects this, and it's very well documented. However, Stuff.objects.filter(rel__a=1).filter(rel__b=1) might not be the same as Stuff.objects.filter(rel__a=1, rel__b=1). This is also very well documented, but in my option this behavior is not always intuitive. Lets take an example: class Tag(models.Model): name = models.CharField(max_length=100) class Entry(models.Model): tags = models.ManyToManyField(Tag) Now if we run Entry.objects.filter(tags__name='stuff') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) WHERE `app_tag`.`name` = 'stuff' If we run Entry.objects.filter(tags__name='stuff').filter(tags__name='other') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) INNER JOIN `app_entry_tags` T4 ON (`app_entry`.`id` = T4.`entry_id`) INNER JOIN `app_tag` T5 ON (T4.`tag_id` = T5.`id`) WHERE (`app_tag`.`name` = 'stuff' AND T5.`name` = 'other') Two JOIN is exactly what we wanted - a WHERE with a single JOIN wouldn't make sense anyway. A different example* Suppose we want to model books that have multiple authors: class Author(models.Model): nationality = models.CharField(max_length=100) sex = models.CharField(max_length=1) birth = models.DateField(max_length=100) alive = models.BooleanField(default=True) class Book(models.Model): authors = models.ManyToManyField(Author) What if we want to get … -
Django sticky queryset filters
In Django, Stuff.objects.filter(a=1).filter(b=1) is almost always the same as Stuff.objects.filter(a=1, b=1). Everyone knows and expects this, and it's very well documented. However, Stuff.objects.filter(rel__a=1).filter(rel__b=1) might not be the same as Stuff.objects.filter(rel__a=1, rel__b=1). This is also very well documented, but in my option this behavior is not always intuitive. Lets take an example: class Tag(models.Model): name = models.CharField(max_length=100) class Entry(models.Model): tags = models.ManyToManyField(Tag) Now if we run Entry.objects.filter(tags__name='stuff') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) WHERE `app_tag`.`name` = 'stuff' If we run Entry.objects.filter(tags__name='stuff').filter(tags__name='other') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) INNER JOIN `app_entry_tags` T4 ON (`app_entry`.`id` = T4.`entry_id`) INNER JOIN `app_tag` T5 ON (T4.`tag_id` = T5.`id`) WHERE (`app_tag`.`name` = 'stuff' AND T5.`name` = 'other') Two JOIN is exactly what we wanted - a WHERE with a single JOIN wouldn't make sense anyway. A different example * Suppose we want to model books that have multiple authors: class Author(models.Model): nationality = models.CharField(max_length=100) sex = models.CharField(max_length=1) birth = models.DateField(max_length=100) alive = models.BooleanField(default=True) class Book(models.Model): authors = models.ManyToManyField(Author) What if we want to … -
Django sticky queryset filters
In Django, Stuff.objects.filter(a=1).filter(b=1) is almost always the same as Stuff.objects.filter(a=1, b=1). Everyone knows and expects this, and it's very well documented. However, Stuff.objects.filter(rel__a=1).filter(rel__b=1) might not be the same as Stuff.objects.filter(rel__a=1, rel__b=1). This is also very well documented, but in my option this behavior is not always intuitive. Lets take an example: class Tag(models.Model): name = models.CharField(max_length=100) class Entry(models.Model): tags = models.ManyToManyField(Tag) Now if we run Entry.objects.filter(tags__name='stuff') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) WHERE `app_tag`.`name` = 'stuff' If we run Entry.objects.filter(tags__name='stuff').filter(tags__name='other') we'd get roughly something like: SELECT `app_entry`.`id` FROM `app_entry` INNER JOIN `app_entry_tags` ON (`app_entry`.`id` = `app_entry_tags`.`entry_id`) INNER JOIN `app_tag` ON (`app_entry_tags`.`tag_id` = `app_tag`.`id`) INNER JOIN `app_entry_tags` T4 ON (`app_entry`.`id` = T4.`entry_id`) INNER JOIN `app_tag` T5 ON (T4.`tag_id` = T5.`id`) WHERE (`app_tag`.`name` = 'stuff' AND T5.`name` = 'other') Two JOIN is exactly what we wanted - a WHERE with a single JOIN wouldn't make sense anyway. A different example * Suppose we want to model books that have multiple authors: class Author(models.Model): nationality = models.CharField(max_length=100) sex = models.CharField(max_length=1) birth = models.DateField(max_length=100) alive = models.BooleanField(default=True) class Book(models.Model): authors = models.ManyToManyField(Author) What if we want to … -
Custom queries III: MySQL
Ok, we saw how to connect to an SQLite database in Python. Now let's see MySQL. There's only a small difference in syntax, but it's significant. The steps, however, are the same.Import the libraryimport MySQLdb as msNote: The as specifies an alternate name for the library. In this case, I won't have to type MySQLdb again and again. I'll just type ms.Establish a connectionconn = ms.connect (host, username, password, database_name)If you're running this on a machine, normally host is localhost. You'll have to create a root user, which you'll have to see how to do in your particular OS on the Internet. Provide that username and password. Then execute the command create database db_name and use this database as the fourth field.Exhausted? So am I. I know it's a lot of effort, but trust me, the results are totally worth it.Create a cursor as usualcur = conn.cursor ()Remember I said %s doesn't work in SQLite? Well, that's the only thing that works here. Let's write a query.query = "select name from employee where empid = %s" % empidREAD THIS VERY CAREFULLY:Now, empid is an integer, but you still have to pass is as a string. Assume empid is 1234. The … -
Django and IPython Notebook
The IPython Notebook is a really cool application, and I've always wanted to use it during Django development and debugging. The only problem is that it requires a lot of dependencies and I feel no need to encumber my production projects with those dependencies for a feature that I ... -
Supervisor with Django and Gunicorn
Supervisor with Django: A starter guide This post assumes that you have used gunicorn and know what it does. I will try everything inside a virtual environment and hope you do the same. What is supervisor. Supervisor is a monitoring tool that can monitor your processes. It can restart the process if the process dies or gets killed for some reason. Use of supervisor: Why I started using it. In production, I use gunicorn as web server. I started a gunicorn process as a daemon and logged out from the server. My site ran as expected for few days. All of a sudden, we started getting '502 Bad Gateway' and I had no idea why. I had to ssh to the server to find out what went wrong. After ps aux | grep gunicorn, I found out gunicorn wasn't running anymore. My gunicorn process died on its own, and I had no idea when and why. Had I used supervisor, supervisor would have been controlling the gunicorn process. It must have recieved a signal when gunicorn died and it would have created a new gunicorn process in such scenario. And my site would have kept running as expected. Other scenario …
