Lots to consider, Robin. Here's some "good" news: that's not some strange problem unique to you, though thankfully it's rare. And there WILL be an explanation and solution: you need not merely suffer the problem.
And yes, there may well be a log explaining it, or something you can enable to understand it.
First, let's clarify: you start saying cf "went down", but then you say you "restarted" it. Did you have to START it (because cf was stopped)? Or did you really restart it, meaning only it was somehow hung? And if you did that, did it restart readily?
If cf truly crashes (goes from running to not running), it will usually write an "hserr" (Hotspot jvm crash) log within the cfusion/bin folder--not where most folks ever consider looking. If you have one there (at the time being considered), read the first few lines of it to see where it reports the cause and tell us.
Assuming there's none or that it did hang rather than crashed, sadly you often won't find an explanation for that. That said, one possibility (which you didn't mention) is the coldfusion-error.log. In there, is there any o "utofmemory" error at the time or just prior to the hangup?
Finally, if none of those bear fruit, in my experience (helping folks solve such problems about weekly for nearly 20 years) the next best step is to implement FusionReactor as a monitor. While it's commercial, there's a free 14-day trial. Besides the very helpful metrics it captures, it also creates valuable logs which nearly always help point directly to the problem.
(There's also the free Adobe PMT tool, which some find to be heavy-handed in its resource usage especially if implemented on the same server as cf, or challenging to implement on a separate server.)
As for what may be amiss, there can be many explanations. For example, I've helped people find it was some single request that when executed (perhaps with unusual form selections) overtook available resources.
A more common problem, especially recently, is an unexpected significant increase in traffic (either in volume or nature). Some folks dismiss that as they presume their measures of traffic show no such increase, but they're being misled.
This was found in another thread here just days ago. And as I anticipated, it was ai bots, search engine crawlers, and other automated requests causing the problem. As for how best to solve that, there are many possible solutions, some easier than others, but they all require careful consideration. It all becomes more clear with the right diagnostics and perspective.
And sometimes we may find that there is a need to modify some configuration settings. It's hard to say from the outside what's "right" for everyone. Again a tool like fr can help in coming up with approprate settings.
And to be clear, you don't have to go this alone, or await/wade through back and forth here. I help people assess (and if needed, implement) such diagnostics to find and resolve such problems about daily (and often in less than hour). If interested, you can find more on my rates, approach, satisfaction guarantee, online calendar, and more at carehart.org/consulting.
Otherwise, besides responding to me--and awaiting possible input from others--there are indeed many resources online that elaborate on these things, from blog posts to presos to forum threads like this. Sadly, there's no single resource gathering those all up. (I created one years ago as cf9f11.com, but it was a lot of work and rarely was referenced. When I removed it, no one complained. We live and learn.)
Anyway, we're here for you. Looking forward to your response. Again, you need not suffer this hangup for long. There will be a resolution, hopefully quickly.
/Charlie (troubleshooter, carehart. org)
1
Reply
AdChoices