In the last sprint we found that the performance issue was caused by a large
number of one-time keys created by crawlers accessing the tracker. We reduced
the time such keys stay active and used a robots.txt to avoid some unnecessary
crawler queries. So far, it seems that this fixed the problem.

I just accessed this issue after creating it
(http://issues.fast-downward.org/issue863) and the first time was very slow as
well even though there was no email involved. Subsequent calls where fast
because the page was probably cached. I made the title a bit broader.

After switching the issue tracker to Docker and upgrading it to the newest
version it is sometimes very slow in loading a page. It seems that this is in
situations where it sends emails. This could be caused by waiting while sending
emails instead of sending them asynchronously.