What to do with 408 timeout?

When trying to access Walrus I lately get a lot of timeouts. This happens when starting an instance (e.g. it's in pending state forever) or when accessing S3 with euca-delete-bundle:

spindler@lysander:~/Work/cloud$ euca-delete-bundle -a $EC2_ACCESS_KEY -s $EC2_SECRET_KEY -U $S3_URL -b image3 -p root.img --clear
-------------------------
         4 0 8
path=/services/Walrus/image3/?max-keys=0
Failure: 408 Request Timeout

-------------------------
-------------------------
         4 0 8
path=/services/Walrus/image3/?max-keys=0
Failure: 408 Request Timeout

-------------------------
-------------------------
         4 0 8
path=/services/Walrus/image3/?max-keys=0
Failure: 408 Request Timeout

-------------------------