Resolved Monitoring does not work with http3

[enerspace]

Username:

TITLE

Monitoring does not work with http3

PRODUCT, VERSION, OPERATING SYSTEM, ARCHITECTURE

Plesk Obsidian 18.0.61 Update #1

PROBLEM DESCRIPTION

If http3:

plesk bin http3_pref --enable -panel -nginx

is activated, monitoring no longer works. The following error is displayed in the console:


monitoring_network?kiosk=tv&theme=light (failed) net :: ERR_SSL_PROTOCOL_ERROR
monitoring_network?kiosk=tv&theme=light (failed) net :: ERR_SSL_PROTOCOL_ERROR
monitoring_cpu?kiosk=tv&theme=light (failed) net :: ERR_SSL_PROTOCOL_ERROR
monitoring_hdd?kiosk=tv&theme=light (failed) net :: ERR_SSL_PROTOCOL_ERROR
X monitoring_cpu?kiosk=tv&theme=light (failed) net :: ERR_SSL_PROTOCOL_ERROR

STEPS TO REPRODUCE

Activate http3:

plesk bin http3_pref --enable -panel -nginx

ACTUAL RESULT

Monitoring not Working

EXPECTED RESULT

Monitoring Working

ANY ADDITIONAL INFORMATION

(DID NOT ANSWER QUESTION)

YOUR EXPECTATIONS FROM PLESK SERVICE TEAM

Confirm bug

 

[Kaspar@Plesk]

Just to clarify the issue a bit better, which monitoring tool are you using? The build-in Monitoring (based on Grafana) or the 360 Monitoring?

 

[Hangover2]

We can confirm this bug. We updated some test servers to Plesk Obsidian 18.0.61 Update #1 too. The build-in monitoring tool is broken and does not display any graphs anymore. The Google Chrome development console shows the errors mentioned in this bug report:

Failed to load resource: net::ERR_SSL_PROTOCOL_ERROR
https://[MY-PLESK-SERVER]:8880/modules/grafana/service/d/monitoring__ram?kiosk=tv&theme=light
[...]

In our case we did not activate HTTP/3 yet.

 

[Kaspar@Plesk]

@Hangover2 are you on Plesk version 18.0.61#1 too?

 

[Hangover2]

@Kaspar@Plesk Yes latest version Plesk Obsidian 18.0.61 Update #1.

 

[ChristophRo]

We've upgraded several servers to 18.0.61 Update #1 (Debian 10, 11 and 12) and so far the internal monitoring tool of Plesk still works on all of them. (we did not enable http3 anywhere)

 

[Hangover2]

@ChristophRo We tested on Debian 11 and 12. Both operating systems are affected for our servers. If you have no problems at all maybe we have a conflicting extension or setting. We will investigate it further.

 

[Kaspar@Plesk]

If been able to replicate the issue with HTTP3 enabled and the issue has been submitted for further investigation with ID: PPS-16028. Thank you for reporting this @enerspace.

@Hangover2 I can't replicate this issue other then with HTTP3 enabled. If you have more information and specific server configuration I'd be happy to check.

 

[Kaspar@Plesk]

Confirmed as a bug with ID: EXTPLESK-5535. Will be shown on the change log page when updated. Unfortunately there is no ETA for a fix release at this time.

Suggested work around is to disabled HTTP3 with plesk bin http3_pref --disable -panel -nginx command.

 

[Longi]

I have the same issue,
Putting the :8443 por in the url makes the monitor work.

https://plesk-server

8443/modules/monitoring/index.php/index/main?context=extNavButton_monitoring_1

Can someone confirm this work around?

 

[Hangover2]

@Kaspar@Plesk

In our case the problem is caused by our login URL to the Plesk interface. We do not use Port 8443 to login to the admin panel.

If we login via this port, the monitoring works like a charm.
e.g. https://[hostname-plesk-server]:8443/

But if we login without this port, the monitoring fails ("Tools & Settings" => "Customizing Plesk URL"):
e.g. https://[hostname-plesk-server/

In the second case the Port 8880 gets added to some URLs leading to the error mentioned.
In Plesk Obsidian 18.0.60 #1 nothing is added.

It looks like the port handling was changed with this release. We have clients that have their own branded login pages to the Plesk admin panel, e.g. "hosting.[client-domain]" (on the same or another server). Some of them are not working anymore with our solution to use a proxy_pass directive. We are still investigating that issue and we will submit an independent bug report soon, if we do not find a workaround.

Both bugs are right now big blockers for an upgrade to the latest Plesk version on our live systems.

 

[Kaspar@Plesk]

Thank you for the additional information, that helps a lot to better identify the root cause of the issue.

 

[manos]

Confirming same bug at 4 almalinux 8.9 Plesk servers (18.0.61)

 

[Hangover2]

The issue also affects other plugins, e.g. the "Google Authenticator". A redirect is triggered including port "8880" leading to the know "ERR_SSL_PROTOCOL_ERROR".

 

[Hangover2]

With Plesk Obsidian 18.0.61 Update #3 the build-in monitoring tool is working again, same as the Multi-Factor Authentication (MFA) extension - at least for HTTP/2.

 

[Hangover2]

Maybe as additional information for the developers of Plesk directly here in the bug report:

- This issue did also disappear in other affected areas (e.g. PHPMyAdmin / [...]).
- Reverse proxying was also affected: we have special domains that are reverse proxying the Plesk admin panel to have individual login URLs for our clients and resellers. This was not working anymore since Plesk 18.0.61, when the proxy_pass URL was different from the individual login URL. With Update 3 for Plesk 18.0.61 the problem was luckily washed away too. We hope it stays this way.

Please take care this issue when you make a new run for HTTP/3. Thanks in advance @Plesk Heroes!