Gap in Data across all monitored servers

[sfereday]

OK, this is a strange one for me. Starting yesterday I'm seeing strange gaps in data. The image below is from the website monitoring, but it seems to extend to other items being monitored as well. Such as CPU usage and RAM usage. All of my servers under latest data have gaps at these times. This wouldn't be an issue, but Zabbix started to alert during these times saying the site was down, when it wasn't. I am having the network looked at as well, but since it's happening with the agent on the zabbix server as well, it makes me think the network may not be the culprit. Any thoughts?

Zabbix server is:
Zabbix Version 5.0.3
Zabbix Agent 5.0.5
CentOS 7.8.2003

[sfereday]

Quick update. Found the one around noon was me rebooting the server.

[sfereday]

I have two different servers monitoring the websites as pictured. Both servers say at 2000 there is no data for these websites. The other items that are monitored, CPU usage, etc, are all still reporting correctly. The setup I have on most of websites is below

[sfereday]

Not sure if anyone has been tracking on this, but I think the issue is write access to the local drive and Zabbix not being able to write quick enough. I've reduced the frequency of items being monitored and put the VM on a disk with higher throughput and the issue seems to have gone away.

[Zdenek_OMNISENSUIT]

Hello.

As I don't know your zabbix environment, I can recommend you to separate db to another vm or at lease use separate fs for your db data and logs (depends on your db type) or add some ram for your db etc.
Use zabbix self-monitoring (zabbix os, db etc) to be able found where exactly your problem is.

-------------------------
omnisensuit.com

[sfereday]

There is a wonky setup here. The throughput issue came up due to slow HDD and DRBD being poorly optimized. I've moved the guest to another host and the issue has pretty much gone away. I'm going to increase the items that are being checked and some of the frequencies to test, but I'm pretty sure it was the host it was on. Throughput to the HDD was about 20MB/s. The new host isn't much better (120MB/s), but it's better than what it was.