A catastrophe based model seems more appropriate. You see big refactoring taking place. Only in a big number sum those Poisson-like events turn into exponentials.
2010 was a major transition for pyramid as it was re-branded from repoze.bfg and merged with the pylons project.