Issue 504 Gateway Time-out since days from a specific subdomain

[PeopleInside]

No, is not that...
Because this error is from Plesk, I get sometimes also in other domains.

 

[Peter Debik]

I see we're already on a second page of this case here on the forum. Normally a 504 error can be easily fixed. Is the error reproducible within a reasonable number of steps? Would you maybe want to consider opening a support ticket on this to let support staff check this directly on your server?

 

[PeopleInside]

I don't have a premium support included.
I dont now steps, this issue comes randomly and it's very strange.

 

[PeopleInside]

Hi, I have reduced this issue by doing some steps:
- increase pm.max_children to 900
- adding in PHP settings Additional directives:

pm.process_idle_timeout = 10s
fastcgi_send_timeout 120s;
fastcgi_read_timeout 120s;
request_terminate_timeout = 300s;

As soon I do this I think I start to receive emails from Plesk that resources was too much used.
I so edited the Monitoring to alert me when resources are used at 80 % and not 20 % as was set by default.

Doing this for now deleted the downtime of my helpdesk subdomain.
This morning I had a 12 min downtime about my community and the alert by Plesk was:

The swap usage status is critical!
The current value is 1.83 GiB.

Any idea to what I can do for resolve this issue? I suppose this time the usage was very height but what can be? An attack?
My VPS seems to be powerful.








The following screenshot is from my webhosting and show you that my VPS resources are not used at 100%

So why I am having downtime and why the swarp memory is so high. I dont know what the swarp memory is

 

[PeopleInside]

I see the swarp file can be something from the operative system.

 

[Peter Debik]

Swap is used when RAM becomes insufficient. It can be a sign that your application needs more RAM, but normally if an application is hungry for RAM, the issue is located in the application, because it should free RAM once it is done doing something. Again this situation is pointing to an infinite loop somewhere, like a nested rewrite or a nested page with missing resources that are creating infinite script iterations. These again lead to an excessive RAM usage. This causes the 504 error, because scripts never end, so they are not responding to Nginx so Nginx displays the 504 error. Back to field no. 1, you will need to check what the scripts of the application are doing.

 

[PeopleInside]

Swap is used when RAM becomes insufficient. It can be a sign that your application needs more RAM, but normally if an application is hungry for RAM, the issue is located in the application, because it should free RAM once it is done doing something. Again this situation is pointing to an infinite loop somewhere, like a nested rewrite or a nested page with missing resources that are creating infinite script iterations. These again lead to an excessive RAM usage. This causes the 504 error, because scripts never end, so they are not responding to Nginx so Nginx displays the 504 error. Back to field no. 1, you will need to check what the scripts of the application are doing.

Thanks. Any suggestion on how check what is the cause of this? How can I understand if the issue come from helpdesk or community and what script or code is causing the issue?
Thank you!