Enhancement idea: show failed jobs on the dashboard until gs-server is restarted

[ethus3h]

trafficstars

As in the title. This would enable easier noticing when something goes wrong like in #52....

[ivan]

This wouldn't help you notice a Segmentation fault because grab-site can't send a message to gs-server when it segfaults. (ArchiveBot has a different design where a parent process is responsible for sending messages, so it can capture the stdout and stderr regardless.)

[ethus3h]

Ah, hmm, oh well. Maybe the gs-server could remember the most recent line of each job, so that stuck ones would become evident?

[ethus3h]

Strike my previous idea; I think it would slow down gs-server probably.... How about, the gs-server could remember the address and identifier of each job?

[ivan]

Remembering the last few events for each job wouldn't slow down gs-server. It's probably not too hard to implement either (just use a Python deque).