Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Alternative to webapp.RequestHandler using memcache for Google App Engine
Fetching latest commit...
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



A small library for Google App Engine to supplant the standard webapp.RequestHandler with a version that will serve pages from memcache where possible. This should improve page response times, and possibly reliability as well (in my personal experience, memcache is much more reliable than the datastore on App Engine).


There's a small sample application showing how the library can be used, but in brief...

  1. 'import mchandler' in .py files that have classes derived from webapp.RequestHandler
  2. Change class definitions from webapp.RequestHandler to mchandler.MemcachablePageHandler
  3. get() methods should be preceded by a @mchandler.memcachable decorator
  4. get() methods should return the rendered content so that it can be cached. You could just return a string, but it's better to return a HeadersAndContent object, as this allows customizations of HTTP response headers etc


Changing webapp.RequestHandler to mchandler.MemcachablePageHandler isn't strictly speaking necessary - you can put a decorator on the former and it should still work. The latter does give you a head() handler, which is nice for reducing spurious errors in the GAE dashboard shif you get robots doing HEAD requests.

If you don't want the content you generate to be cached, just return None at the end of your handler. will try to import a module cachability containing the function can_use_cached_copy(), to determine if it is safe to cache a page. The default functionality is fairly basic, and the file/function should be extended to cover whatever your requirements are. Of course, if you've got a handler that's producing content that can't be cached most of the time, then all of this is a bit of a waste of your time...

This code was originally part of my blogging system, and there are calls to code that does some internal analytics type stuff e.g. to ensure that pages that get served from the cache also get logged. This library should work fine without that (just calls a dummy stub instead), and if you're using something like Google Analytics you'll be fine. (As I use the Ghostery Firefox extension whilst browsing, it felt a bit hypocritical to put Google Analytics on my own site :-)

There are a couple of settings you can tweak if desired, have a look at the comments at the top of


As of 2010/12/21 I'm not aware of any, but this code is still pretty new and basic.

The main thing is that you need to be super-careful with what gets cached and what doesn't.


I haven't tried it personally, but this looks to cover similar ground:


Written by John Smith December 2010 My blog (which uses this code): My Twitter: John MMIX


GPL v2 -

Something went wrong with that request. Please try again.