Indeed, I was just coming back to respond and saw this post. Right after my previous post, it occurred to me to do this, and that worked.
I’ve also managed to resolve the primary issue.
Edit: I just now came back to double-check something and saw that you had also mentioned this next part in your previous post; I guess that shows just how close we were and what a good night’s sleep can do! 
Because I use index lifecycle management, I have index auto-creation disabled, with the following exceptions :
{
"persistent" : {
"action" : {
"auto_create_index" : ".*,sg7-*,searchguard,searchguard*,ilm-history-*"
}
}
}
This allows system indices matching the following pattern to be automatically-created: .*. However, date math indices start with <, so I had to make the following update to this setting:
{
"persistent" : {
"action" : {
"auto_create_index" : ".*,<.*,sg7-*,searchguard,searchguard*,ilm-history-*"
}
}
}
After making these two changes, everything works as expected. The .signals_log_* index is automatically created, and I can see my execution history:
The default timestamped Elastic Stack indices (i.e., the .monitoring-* indices) don’t have this issue, although I found out the hard way that they require auto-creation for .* to be enabled. I suspect that the date math must be handled on the client side for these indices.
If I may suggest, you may want to add the following information to the documentation:
- URL encoding is not required for the date math in logging indices.
- Enable
<.*index exception if index autocreation is disabled.