- 2015-08-21: 50% less server load with MongoDB on the Drupal 7 site factory at France Télévisions
- 2015-07-15: Our first Drupal 8 production site at France Télévisions is live
- 2014-08-18: 400% speedup in 3 weeks for http://france3-regions.francetvinfo.fr/ : who said Drupal back-offices had to be slow ?
- 2014-02-07: Sotchi Olympics traffic not a problem for http://www.francetvsport.fr/ , which I rearchitected on Drupal 7 in 2013
- 2011-09-14: Completed migration of FranceInfo.FR from SPIP to Drupal
- 2011-07-13: The new social network features of Le Figaro are now powered by an OSInet-designed MongoDB implementation
- 2010-12-21: Madame Figaro brand new site by OSInet and others
- 2010-08-16: France.FR is back online with OSInet and Typhon
- 2010-06-15: the new France Culture, which OSInet helped reach its performance goals, is now online
A few days ago, while I was writing a bit of Silex code and grumbling at Doctrine DBAL's lack of support for a SQL Merge operation, I wondered if it wouldn't be possible to use DBTNG without the rest of Drupal.
Obviously, although DBTNG is described as having been designed for standalone use:
DBTNG should be a stand-alone library with no external dependences other than PHP 5.2 and the PDO database library, in actual use, the Github DBTNG repo has seen no commit in the last 3 years, and the D8 version is still not a Drupal 8 "Component" (i.e. decoupled code), but still a plain library with Drupal dependencies. How would it fare on its own ? Let's give it a try...
WebProfilerServiceProvider brings the Symfony Profiler to Silex apps, and with it the nice "Timeline" feature, detailing for each request cycle the time spent handling each event dispatched by the
kernel.terminate, with detail time spent in each individual listener.
The timeline looks like this one:
This is all nice and well, but there is a problem with this timeline: although the beginning and end of the request cycle is profiled in depth, it represents less than half of the total time spent for this request.
More than half of the total request time is spent in the controller, and this example does not even perform any database access, which would make the controller represent event more of the total time, making the detailed analysis of the beginning and end of the request cycle less relevant.
What would make sense would be to subdivide the time spent in the controller in much the same way the time spent in the event listeners is already plotted. Is there a simple way ? Sure: let us see how the profiler works.