I finally got around to hunting down and squashing a bug on my Site Health page. It was one of those bugs that wasn’t easy to reproduce, but I happened to catch it occurring repeatedly on Sunday.
One of the benefits of having C++ code using Wt for this kind of page: I can attach to a running process with the debugger (gdb) and debug a multithreaded process live. The problem was the allocation of an extra data series in the filesystem plot, without assigning header data to that data series. The header data is needed when rendering the legend. If it’s not there… one of the threads either causes a seg fault or hangs.
I fixed the problem and the Site Health page is working again. I will need to update Randy’s server when he is connected at his new location.