Issue Watchdog down after Plesk update

[H.W.B]

Server operating system version

Ubuntu 22.04.5 LTS

Plesk version and microupdate number

Plesk Obsidian 18.0.79 Update #1 Web Host Edition

Hello,
Every time there is an update for Plesk, the watchdog module is switched off (or not turned on).

I noticed that with every new version for Plesk, the watchdoig module is off. This is from version 18.0.75 to 18.0.76, and so on until the latest version 18.0.79.

I every time need to manually switch it on after the update.

Henk

 

[Sebahat.hadzhi]

Hi, Henk. I was unable to replicate the behavior on an Ubuntu 22 server after a Plesk Obsidian 18.0.78 > 18.0.79 update. So, my assumption is that there is a local server issue that is preventing the service from starting automatically. Could you please check the /var/log/plesk/modules/wdcollect.log and let us know what errors/warnings were logged after the Plesk update?

 

[H.W.B]

Hello Sebahat,

This is a part of the log-file

time=2026-07-07T12:32:32.607+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:41892: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:32:32.857+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:41898: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:32:33.108+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:41910: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:32:33.359+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:41922: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:39:15.208+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:50226: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:39:15.459+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:50230: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:39:15.710+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:50232: client sent an HTTP request to an HTTPS server"
time=2026-07-07T12:39:15.961+02:00 level=INFO msg="http: TLS handshake error from 127.0.0.1:50240: client sent an HTTP request to an HTTPS server"

That is what there is.

Henk

 

[Sebahat.hadzhi]

Thanks. That shouldn't be preventing the service from starting. Is that all being logged?

 

[H.W.B]

Hello,

These lines are the only one in the log file. Only the date and times are different. But nothing else....

Henk

 

[Sebahat.hadzhi]

I see. Thank you for the confirmation. How about /var/log/syslog?

 

[H.W.B]

Hello,

This is the Syslog from 07-07-2026

Jul 7 12:32:45 plesk rsyslogd: message repeated 17 times: [[origin software="rsyslogd" swVersion="8.2112.0" x-pid="96118" x-info="rsyslog – the rocket-fast system for log processing pipelines"] rsyslogd was HUPed]
Jul 7 12:32:45 plesk rsyslogd: [origin software="rsyslogd" swVersion="8.2112.0" x-pid="96118" x-info="rsyslog – the rocket-fast system for log processing pipelines"] exiting on signal 15.
Jul 7 12:32:47 plesk rsyslogd: imuxsock: Acquired UNIX socket '/run/systemd/journal/syslog' (fd 3) from systemd. [v8.2112.0]
Jul 7 12:32:47 plesk rsyslogd: rsyslogd's groupid changed to 113
Jul 7 12:32:47 plesk rsyslogd: rsyslogd's userid changed to 107
Jul 7 12:32:47 plesk rsyslogd: [origin software="rsyslogd" swVersion="8.2112.0" x-pid="4159450" x-info="rsyslog – the rocket-fast system for log processing pipelines"] start

Henk

 

[Sebahat.hadzhi]

Thank you, Henk. I cannot precisely identify what's the reason for the Watchdog service's failure to start automatically. I would recommend getting in touch with Plesk support, so further investigation could be conducted directly on the server.

The only workaround I can suggest is to adjust the wdcollect service to restart in 1 second, instead of 100ms (by default) and disable the systemd rapid restart protection for this service.

1. Use the systemd unit editor to create unit overrides config:

systemctl edit wdcollect.service

2. Add the following:

[Unit]
StartLimitIntervalSec=0
StartLimitBurst=1

[Service]
RestartSec=1

3. Reload configuration changes:

systemctl daemon-reload