How to reduce the gluttony of Apache?

Question

Apache HTTP Server

Good afternoon. I'm a fairly layman in setting up servers, learning from the manuals and trying to do it carefully.
The essence of the problem:
there is a VDS, a fairly powerful
6 X Intel (R) Xeon (R) Silver 4114 CPU @ 2.20GHz
6GB of RAM
100GB of SSD
The server has an opencart store with 200 thousand products, there are no more than 10-20 users online at the same time. Daily attendance, God forbid, 100 people a day.
In this situation, everything should fly BUT
Periodically, the server goes down, according to the logs, I found out that at some point the processes exceed 500 (in the normal state 130-150), almost all of this is Apache2 on behalf of the user. Then the memory ends, all the processes and hello rush. Only reboot saves
I didn’t understand where these Apache processes come from, I didn’t notice a surge in online activity.
The muscle base eats a maximum of 500MB out of 6GB
In the normal state, Apache on behalf of the user eats a maximum of 15MB of memory, at the moment when everything falls down - this bunch of processes eats 100-150MB each
Where to dig and what to do?
First of all, why does the Apache eat so much? 100-150MB per process
Secondly, how to get rid of the situation with a memory leak in general?

Answer 1

Daily attendance, God forbid, 100 people a day.

And a legion of bots
To begin with, reduce MaxClients, then dig towards access logs and look for what and why eats memory and clogs the process pool.
Memory eats not apache, but PHP.
In general, no way.

Answer 2

replace with nginx?

Answer 3

I did not notice a surge in online activity.

do you have traffic monitoring?

Answer 4

The solution was, you won't believe it, stupid memcached setup. Allocated 1 GB of memory.
Since there were more than 15 thousand session files in the cache folder of the store, it somehow affected the campaign. As well as what is the dependence of xs, but now more than 800 MB of RAM is not consumed in principle.
Limits in Apache did not touch