Resolved Site goes down EVERY TEN MINUTES

[markmcde]

My site, www.markmcdermott.com (markmcde.x10host.com/wp22/) is being monitored by Hyperping. For the past four, days, I've been getting emails that say the account is "down." A few minutes later, I get a notice that the account is "up." This has been happening practical every ten minutes. And it is affecting access to the site.

WHAT IS GOING ON?
Sample:
Your domain http://www.markmcdermott.com has been down for 5 mins

Date: Wed, 23 Sep 2020 21:03:44 UTC
Status Code: 500 Internal Server Error in New York City

Double checks
500 Internal Server Error in San Francisco
500 Internal Server Error in Frankfurt


See more details: https://app.hyperping.io/5753

 

[garrettroyce]

I'm not seeing a lot of information on what they check or how they're checking it. For example, sending an IMCP echo command to the server (a "ping") may get a reply or it may not. If it does get a reply, does it mean HTTP is working? Or, maybe the server doesn't allow pings and HTTP is working. Or maybe the server blocks pings if they come in too often.

I see a 500 status... is that the return status of your website? A 500 error is logged in DirectAdmin; it should say the origin of that error. 500 doesn't mean that the server is down, it means the server couldn't handle what the request was. This could be a problem in the request to the server or a problem with the piece of software or script handling the reply. If HTTP was down, you would get a time out or connection refused.

But, maybe SQL is down? That could be a 500 error. Or, maybe just a white page. Or, does hyperping test that too?

The problem with these automated uptime services is that they don't give you all of the data you need to determine what's going on. In the past, x10 used to highly discourage and even block them. They give a lot of false positives for down time.

That being said, what do you see on your website? What's in the error logs? If normal requests are working, then maybe hyperping needs to be configured differently.

 

[markmcde]

Thanks for writing. Frankly, the main WordPress page has not given me any issues. I also have a little script on the host that takes Google Calendar data and converts it to HTML, which then feeds it to another blog post inside of iframes. The end result is here:
http://www.chicagonow.com/the-beeronaut/2020/09/chicago-beer-weekend-september-25-27/
While composing that post, I was getting blanks in the individual iframes. It does look as though these aren't coming up on the published page as of yet. I've asked readers to let me know if this becomes an issue.
Yes, it could be that perhaps the free pinging site has changed something in its automation that's causing the messages. I thought I would work backward and see if maybe x10 had made any changes or was migrating again to cause it.

I'm not seeing a lot of information on what they check or how they're checking it. For example, sending an IMCP echo command to the server (a "ping") may get a reply or it may not. If it does get a reply, does it mean HTTP is working? Or, maybe the server doesn't allow pings and HTTP is working. Or maybe the server blocks pings if they come in too often.

I see a 500 status... is that the return status of your website? A 500 error is logged in DirectAdmin; it should say the origin of that error. 500 doesn't mean that the server is down, it means the server couldn't handle what the request was. This could be a problem in the request to the server or a problem with the piece of software or script handling the reply. If HTTP was down, you would get a time out or connection refused.

But, maybe SQL is down? That could be a 500 error. Or, maybe just a white page. Or, does hyperping test that too?

The problem with these automated uptime services is that they don't give you all of the data you need to determine what's going on. In the past, x10 used to highly discourage and even block them. They give a lot of false positives for down time.

That being said, what do you see on your website? What's in the error logs? If normal requests are working, then maybe hyperping needs to be configured differently.

 

[garrettroyce]

Ok. If you find something, let us know. Definitely skim the error logs quickly for anything abnormal. I don't have a recommendation on a monitoring service because I've had so much trouble with them. And, writing something in a cron job would not be accurate because more than likely the server will accept connections from itself, even if outside requests aren't accepted. So, unless you were to write a more comprehensive script to test the several units that make up a functional website, I think the monitoring service is going to be... well a disservice.