Zabbix server health

[tmoises25]

Hi guys, I'm a new member here and need help with my Zabbix Server.
I have notification configured to send messages for my Telegram chat. I recently received the following alert:

Problem: sda: Disk read/write request responses are too high (read > 20 ms for 15m or write > 20 ms for 15m)
Problem started at 08:02:54 on 2023.07.25 Problem name: sda: Disk read/write request responses are too high
(read > 20 ms for 15m or write > 20 ms for 15m) Host: Zabbix server Severity:
Warning Operational data: 0 ms, 62.76 ms Original problem ID: 1987464

My Zabbix Dash


The disk waiting time grow… Then I look at my processor, RAM, and network interface, everything looks good.


Has anyone ever experienced this?

[MRedbourne]

I'm only seeing about an hours worth of logs here. Has this only happened once? If so, I wouldn't be overly concerned, yet. I find it extremely unlikely (read: impossible) that you genuinely hit a 10,000 second (1,000,000ms) read time delay without having some other form of performance degradation on your disk.

I'm not sure of your architecture, but here are some broad questions:

1. Can you post your server specs here (# of CPU Cores, Processor Information (Eg: Xeon Platinum 8268), RAM, # of Disks, Disk Sizes, Disk Types)
2. Are you artificially capped on your disk R/W (either in IOPS or MB/s)?
   1. (More of a concern if you're using Cloud infrastructure like AWS, Azure, etc.
3. If the server is hosted in the cloud by a provider, can you provide Sku information?
4. Zabbix Server Architecture
   1. Eg: Are you running this "all-in-one" (Zabbix Server + Web Front End + DB), or have you split the roles, and in what capacity?
5. Where is Zabbix storing it's information if you're running an all-in-one?
6. Post the Zabbix server version.

Our server architecture right now is a 2-tiered application build. We have 2 Zabbix Server nodes running native HA (Act/Pass). Each of the "server nodes" hosts zabbix-server and the web front end. The backend database is an Azure-hosted MySQL database set for autoscale for both IOPS and storage size.

[tmoises25]

Hi, thanks for your reply.
1) My server specs:


2) No.
3) My server is on premise.
4) Exactly, Zabbix server + front + db. I followed the Zabbix Install guide: https://www.zabbix.com/br/download
5) I'm running Zabbix server in a VM.
6) Sure:

[MRedbourne]

Looks fine, maybe a little memory starved from minimum required specs. (Though, if it hasn't thrown memory alerts yet, I wouldn't worry about that either). 2 Requirements (zabbix.com) I wouldn't worry about the disk warning unless it keeps happening.

The only suggestions I would have is to move the storage of the Zabbix DB onto a dedicated disk or server. Depending on the partitioning of the OS, you may stand to fill the root partition which can cause issues with Linux. The other one being patching - Zabbix Server 6.0.19 is out for LTS. Version 6.0.5 is over a year behind on patches.

[tmoises25]

Thanks for your suggestions! I'll try to fix it with the service update.
Currently, I can't move the storage onto a dedicated disk, so maybe in the future I do it.