Very urgent help! The site and CRM system have been down for more than half an hour, unavailable

At the same time, there is another topic where a similar question was raised. At the time of resource unavailability, such a picture occurred on the server with the database https://prnt.sc/LOOBmUJlEdRl. Usually the problem occurs with a minute crown. and when the load increases, it works very long or cannot work at all, like today.

Website and box https://chekhol.com.ua/

Good afternoon. In order for there to be a solution, you need to understand what is causing the load. Now your system is working normally.

Yes, the time has been prolonged How can you find out? What can I do during the load to find out the reason and pass it on to you?

[quote] Dmitry Savchuk wrote: At the time of resource unavailability, such a picture occurred on the server with the database [file]18082[/file] Usually the problem occurs with a minute crown. and when the load increases, it works very long or cannot work at all, like today. [/quote] Judging by what you have shown here, crowns have nothing to do with it. And it works for a long time, because the load on the database is big. Melon requests are created by external connections, and I would recommend you, at times of load, to pay attention to external connections. Most likely, these are attacks or overtly aggressive behavior of index "bots".

If at any moment the name/availability of 10-20 products is changed at the same time, the system will fail. In a minute, the processor [file]18083[/file] flies up and slows down the entire resource. Maybe it's not the problem, but the brakes are still huge. Periodically, it can give a 502 error and all this only when editing, placing new goods, as well as when confirming orders and reserving the availability of goods in stock.

We get such a situation every day, several times a day [file]18086[/file]. At the same time, we try not to place orders for statuses where goods are reserved often, once every 15-20 minutes. But still, sometimes we get such a story. If the orders are carried out in stages, in which the reserve is carried out upon confirmation, that is, every few minutes, then the resource is constantly available. The processor is also overloaded when orders are received. It is often almost impossible to work.

such a request [file]18107[/file] often appears in SQL We always see such a request when the load is increased

Cloudflare was connected Now there is a higher load on the server The cloud shows that in 30 minutes there were 11,000 calls to voip.php https://prnt.sc/MHRAL5WYM41q. Could this be the reason for the load?