@ask, seems that the respective overall score of servers isn’t being properly updated since earlier today. I have some servers where the overall score doesn’t match the median derived from individual active monitor scores, among them one server that should be added to the DNS rotation again after recovering from networking issues, and another that should be dropped due to networking issues.
I found some servers where the score was near 10 and from my point of view their scores and pool inclusion seemed to work fine. Can you share some more details about the issue you’re facing?
Look for the most recent “recent_median” entry in your servers’ CSV logs as one symptom.
That’s a bug….for sure 
I think it’s just some service not having come up properly again after the database trouble and recovery earlier today, or something got stuck and needs to get unstuck.
1 Like
yeah, the scorer is cautious to not “miss anything”. I think there’s some logic bug in how it processes the scores so if there’s a break in scores available it can fail calculating a median and it bails out. I can’t quite figure out why it hasn’t happened before like this, but I made a tweak so it’ll skip processing a score if it’s sufficiently old (older than 3 hours); and it’ll log some of the details to hopefully figure out what happened.
1 Like
Looks like the system caught up and is back to normal as of ~12-15 hours ago.
I’ve added to the todo some improvements to go into an optimized mode if this particular queue/calculation gets behind.
1 Like