Are you using opcache?

Same issue, a few more details.

Windows 2012/IIS/Plesk Onyx 17.8.11

Several websites with no recent changes, only change is plesk updating PHP 7.2 to 7.2.19 https://docs.plesk.com/release-notes/onyx/change-log/#contents-17811-mu55

Yes it is using opcache.

Yes, opcache is enabled.

I have made some tests and the problem is on with or without the opcache enabled for the subdomain.
The subdomain (cnr.cronobandeira.com) that originally had the problem (with php 7.2.19) is now running php 7.1.30 without any issues.

PHP error log:
PHP Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 7810770492979178392 bytes) ... (That's a lot of bytes!!!, Much more than in the other reports posted).

Possibly related to https://bugs.php.net/bug.php?id=78103 (as posted by mario on that report).

Same issue with PHP 7.2.19 (FastCGI), Windows 2016 server and Plesk.

very similar experience!
two Windows 2008R2 x64 servers
Under Zend Server 2018.0.3 with php 7.2.15 32bit, opcache enabled, no errors, very stable.

Upgrade to php 7.2.20 32bit (no more Zend Server), still with opcache enabled, on same two windows servers.
 
After production server runs OK for ~10-20 minutes in load balancer, I get errors from pages that connect to remote servers, maybe 1% of requests.
Most common failures...
-	Uncaught Exception cURL error 6: getaddrinfo() thread failed to start (connect to remote queue service)

-	[Microsoft][ODBC Driver 17 for SQL Server]SQL Server Network Interfaces: Not enough storage is available to process this command.

-	fsockopen(): unable to connect to searchengine.local:8080 (An operation on a socket could not be performed because the system lacked sufficient buffer space or because a queue was full.

When I disable opcache in php 7.2.20, errors stop, but pages are too slow!

Tried these variety of opcache directives

ONE 
	opcache.enable=1

TWO
	opcache.optimization_level=0xffffffff

THREE
	opcache.optimization_level=0xffffffff
	opcache.mmap_base=0x20000000

FOUR
	opcache.mmap_base=0x20000000

When opcache enabled, with all directives tried, random errors start on server in 10-20 minutes

update from my prior report...
php 7.2.21 64 bit stable in my environment

Steps to resolve
ONE switch to 64 bit php from 32bit
TWO disable extension wincache 2.0.0.8, which never had an official build for php 7.2
THREE use mmap_base as mentioned online for Windows 
FOUR restart Windows 

A windows server restart was needed once these were applied to php.ini. Even App Pool recycle resulted in lingering failures (I have 5 app pools with php).

These directives related to opcache in php.ini now
opcache.error_log = "D:\php72x64\tmp\opcache.log" 
opcache.file_cache="D:\php72x64\tmp\opcache" 
opcache.log_verbosity_level=2
opcache.mmap_base=0x20000000

Bonus, during unstable trial and error time frame, having the file_cache configured prevented a 500 status, instead logging "Sep 10 16:19:22 2019 (6692): Warning Base address marks unusable memory region (fall-back to file cache)"

mmap_base and the removal of wincache seemed to be the primary fix, may try wincache again in future php build, if MS releases official extension

I assume that this is a duplicate of bug #78103, which apparently
is fixed as of PHP 7.2.20, or can anybody still reproduce this
issue?

No feedback was provided. The bug is being suspended because
we assume that you are no longer experiencing the problem.
If this is not the case and you are able to provide the
information that was requested earlier, please do so and
change the status of the bug back to "Re-Opened". Thank you.