ZABBIX Forums - Forum

Zabbix 7.0.26 - 7.0.29 housekeeper consumes all available RAM

MelvinS — Mon, 03 Aug 2026 06:21:19 GMT

Hello,

I have a memory problem with the housekeeper in Zabbix 7.0.26 - 7.0.29 using MariaDB.

Housekeeping was disabled for some time, so old history data accumulated. After starting it manually, the housekeeper continuously consumes more RAM while processing old history and trends:
1281 11303284 11896316 /usr/sbin/zabbix_server: housekeeper [removing old history and trends]
1347 7080 599644 /usr/sbin/zabbix_server: trigger housekeeper [deleted 0 problems records in 0.013036 sec, idle for 60 second(s)]

The VM originally had 8 GB RAM and became unresponsive. I increased it to 16 GB, but the housekeeper consumed that memory as well. More RAM only delays the problem. After restarting zabbix-server, the memory is released and the server remains stable until housekeeping is started again.

This seems related to the initial history queries that retrieve itemid and MIN(clock) from all history and trend tables.

Is this a known issue in Zabbix 7.0? Is there a way to limit the housekeeper’s memory usage or safely process the accumulated history in smaller batches?

Environment:

Zabbix 7.0.29
Debian 13
MariaDB
16 GB RAM

At the moment, the database continues to grow because new history data is constantly being added, while the housekeeper cannot complete its cleanup process without exhausting all available memory.

Best regards

parameters on an action script at host discovery time with actions and mediatype

jhoogie — Sun, 02 Aug 2026 10:18:44 GMT

Hello,

I am a but struggling with my setup.
My goal is to discover network devices on a subnet and based on the type create an item for it in Zabbix with the template etc and also register the device in PHPIPAM
That first part works great.
Just create the discovery rule and bind it to a discovery action.
The hostname is based on the data I retrieved with the snmp querie and fetching .1.3.6.1.2.1.1.5.0 from the device which is the hostname
But now the fun part.
I want to also populate my external IPAM with at least the ip address and hostname
It seems that using a mediatype and call my pythonscript should work.
The mediatype currently has "phpipam_add_host.py" as script with {DISCOVERY.DEVICE.IPADDRESS} {DISCOVERY.DEVICE.DNS}

That works, but only when reverse dns is populated for this device, if there is no reverse dns record than the {DISCOVERY.DEVICE.DNS} is empty
In IPAM I do get the ip address registered, but no hostname
I am facing the challenge that I want to use the value read with .1.3.6.1.2.1.1.5.0 as discovery item for the hostname, but I do not see what parameter value could be used.
From what I understand there is no such thing as {hostname} as the host is not yet created completely in Zabbix.
Any ideas on how to set this up?

John

Agent2 NVIDIA plugin for Windows 7.4.11+ is not a valid Win32 application

evzen — Sat, 01 Aug 2026 19:33:21 GMT

There seems to be an issue - or at least something weirdly changed - in the Agent2 NVIDIA plugin for Windows 7.4.11 and newer.
The plugin does not run, with OS complaining that it cannot create the process because it's not a valid Win32 application.
All versions up to 7.4.10 work perfectly fine.

As a proof I collected the plugins 7.4.8 to 7.4.13 and one thing is clearly visible right away - the executable sizes changed significantly since 7.4.11:

Code:

C:\Program Files\Zabbix Agent 2>Directory of C:\Program Files\Zabbix Agent 2

13.03.2026  17:50        10 702 192 zabbix-agent2-plugin-nvidia-gpu-7.4.8.exe
09.04.2026  10:31        10 702 192 zabbix-agent2-plugin-nvidia-gpu-7.4.9.exe
28.05.2026  13:00        10 702 704 zabbix-agent2-plugin-nvidia-gpu-7.4.10.exe
02.06.2026  18:19         8 602 696 zabbix-agent2-plugin-nvidia-gpu-7.4.11.exe
09.07.2026  11:56         8 602 696 zabbix-agent2-plugin-nvidia-gpu-7.4.12.exe
29.07.2026  13:42         8 745 528 zabbix-agent2-plugin-nvidia-gpu-7.4.13.exe

Trying to execute the plugin - version 7.4.10 works:

Code:

C:\Program Files\Zabbix Agent 2>zabbix-agent2-plugin-nvidia-gpu-7.4.10.exe
NVIDIA plugin for Zabbix agent 2

Usage of "zabbix-agent2-plugin-nvidia-gpu-7.4.10.exe"
    zabbix-agent2-plugin-nvidia-gpu-7.4.10.exe -h
    zabbix-agent2-plugin-nvidia-gpu-7.4.10.exe -V

Options:
  -h --help                               Display this help message
  -V --version                            Print program version and exit

...but version 7.4.11 or newer does not launch:

Code:

PS C:\Program Files\Zabbix Agent 2> .\zabbix-agent2-plugin-nvidia-gpu-7.4.11.exe
ResourceUnavailable: Program 'zabbix-agent2-plugin-nvidia-gpu-7.4.11.exe' failed to run: An error occurred trying to start process 'C:\Program Files\Zabbix Agent 2\zabbix-agent2-plugin-nvidia-gpu-7.4.11.exe' with working directory 'C:\Program Files\Zabbix Agent 2'. The specified executable is not a valid application for this OS platform.At line:1 char:1
+ .\zabbix-agent2-plugin-nvidia-gpu-7.4.11.exe
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.

First pass at a DB2 by ODBC template

jtnfoley — Fri, 31 Jul 2026 17:42:25 GMT

I spent a morning prompting my way thru a Google Gemini session, generating a zabbix 7.4 template with discovery via item, trigger and graph prototypes.
The items being solicited were also pulled from Gemini, prompting for key health and performance metrics that are accessible by SQL.
ODBC DSN, username, password, and a CSV list of databases (schemas?) are the four macros.
DB2CLI and DSDriver from 11.1 (my youngest target server) was installed on the Zabbix host and proxy.
/etc/odbc.ini and /opt/ibm/dsdriver/cfg/db2cli.ini have the ODBC DSN. The macro username exists on the target Linux DB2 servers as a normal (/sbin/nologin) user which is a member of the dbmonitor OS group.

I have a single DSN for one schema on a given server, the others in the CSV macro are accessible via the named schema (queries to the other schemas are not fenced by the connection named in the ODBC DSN.)

One of the discovered items is the HADR state. I'm not doing anything with it currently, as I'm trying to figure out how to disable polling of values on the HADR "Standby" hosts, polling only the HADR "Active" (or HADR null?) host.
Note that this is VERY young - my DBAs haven't even blessed the queries and metrics being polled. I figured there's enough unanswered requests for DB2 templates, this can be a beginning for the rest of the community.

Any macro with "NOTE: Escalation cancelled ..." in?

troffasky — Fri, 31 Jul 2026 15:59:20 GMT

Have a custom webhook sending a JSON event for trigger fire/clear. I have noticed that for an "escalation cancelled" event, this bit of the message does not appear in our payload and I cannot work out what macro it would be in. So it just looks like the original problem webhook being sent an hour later.
I have tried {ALERT.MESSAGE} but that is just the original message without the 'NOTE: Escalation canceled: trigger "" disabled.' part.
Is there any macro I could use to infer that the trigger was disabled?
Zabbix 7.0

Veeam B&R template requires service account with backup admin privileges

ezoltan — Fri, 31 Jul 2026 11:54:56 GMT

Hi,

Hope this will save some time to those struggling to make the Veeam B&R template work with Veeam B&R 13.

I've just installed the Veeam template from https://git.zabbix.com/projects/ZBX/...ads%2Frelease% 2F6.0. It looks like it hasn't been updated since 22/01/2024, well over 2.5 years.

I am on VBR 13.1.0.411, Zabbix Server 7.4.0 and a bunch of Zabbix Proxies ranging from 7.4.0 to 7.4.13 running on Ubuntu 24.04.LTS.

The instructions state:

Setup
1. Create a user to monitor the service or use an existing read-only account.

I found that a read-only account will be denied access and results in the following error:

What fixed it for me is giving the user Backup Administrator in Veeam's Users and Roles dialog.

If the permissions necessary to query certain metrics are higher than Backup Viewer / read-only access then the documentation is incorrect and should be updated to reflect it. Otherwise the bug should be fixed. Veeam 13 has been out for quite some time.

Did not fiddle with various permission levels.

Regards,
Zoltan

дашборд с кнопкой

Hell Berrel — Thu, 30 Jul 2026 08:36:06 GMT

Приветствую всех
zabbix 7.4
Делаю мониторинг контроллера солнечных батарей. Данные с него передаются через агента2 протоколом modbus. Данные о текущем состоянии идут нормально. Есть пара регистров, которые нужно записать в него. Мне надо раз в полгода поменять значение battery.type и battery.capacity и записать новые значения. Потом через полгода поменять обратно.

В принципе, я с помощью скриптов на питоне (выполнение определенной команды с параметрами) получилось записать в реестр на контроллере. Есть желание сделать дашборд типа такого,
где в выпадающем списке выбираю тип батареи, а в поле емкости надо ввести значение и нажать на кнопку. Значения должны записаться и при обновлении виджета отобразить текущее состояние. Такое вообще возможно?
Сам не программист, замучил chatgpt, подсовывал ему и официальную документацию и готовые модули и дашборды, кружит кругами меня.
Чего удалось добиться, это на примере из документации сделать такое

при вводе какого-нибудь значения "текст" в поле и нажатии на кнопку внизу появляется надпись Hello "текст"
И опять он собирирается меня по третьему кругу завернуть

zabbix configuration for port 443 on nginx,mysql,php raspberry pi

colporteur — Wed, 29 Jul 2026 21:35:51 GMT

How to configure zabbix on an 8G Raspberry pi 5 OS (13 (trixie)), nginx (1.26.3-3+deb13u7), mysql (5.8+1.1.1), php-mysql (2:8.4+101) & zabbix (1:7.4.13-1+debian13) to support https using self-signed-certificates.

I have installed nginx and have it working on port 433 using the self-signed-certificates. I then installed php and it serves webpage on 443. I then installed mysql and finally zabbix. I can connect to the zabbix web interface console :443 but the error message at the bottom indicates Zabbix Server is not running:the information displayed may not be current.

Before starting the project to use port 443, I built a zabbix server using the same configuration on port 8080. The install is working. I have configured a few devices and am gathering data.

I am a new user to zabbix. No experience at all! After evaluating a few products I landed here. The goal is to replace the MikroTik Dude server with zabbix. If someone could share their knowledge or point me in the direction of https configuration resources, that would be appreciated.

So far I have tried internet searches and cull this forum looking for hints. After to many attempts to continue to count, I ask for help.

I’m close to making it work, just not close enough.

Zabbix 7.4.13 stable version is out now!

Michael_Kammer — Wed, 29 Jul 2026 10:59:59 GMT

The latest Zabbix 7.4.13 stable version has been released!

Have a look at the new features and improvements:

7.4.13: https://www.zabbix.com/rn/rn7.4.13

All new Zabbix releases can be downloaded here: https://www.zabbix.com/download

Zabbix 6.0.48 and 7.0.29 stable versions have been released!

Michael_Kammer — Tue, 28 Jul 2026 13:03:27 GMT

The latest Zabbix 6.0.48 and 7.0.29 stable versions are out now!

Take a look at the new features and improvements:

6.0.48: https://www.zabbix.com/rn/rn6.0.48

7.0.29: https://www.zabbix.com/rn/rn7.0.29

All new Zabbix releases can be downloaded here: https://www.zabbix.com/download

SNMP and different communities on one device

markosa — Tue, 28 Jul 2026 11:28:12 GMT

Hi everyone.
Has anyone stumbled across this and resolved it without external scripts, one mgmt IP responds differently based on used community, like some stacked switches do.
I need to do lld to device which is monitored using community1 and then on same device i need to do lld using community2, is that even possible? I haven't figured that out yet without adding new interface(same IP) and assigning different community to that and then assign template items to use that interface, this is quite painful way...

Оповещения, замена не работающему telegram

Sunrise_94 — Tue, 28 Jul 2026 11:02:30 GMT

Здравствуйте

У меня настроены оповещения в telegram + email, поскольку нынче telegram не работает на просторах России, прошу поделиться другими решениями, достойными telegram(max не подходит). Либо, если есть такая возможность, то как настроить некий прокси, чтоб сервер смог продолжать высылать оповещения на telegram.

Передача названия процесса, найденного LLD, в скрипт в действия триггера.

Victor Sklyarov — Tue, 28 Jul 2026 09:55:57 GMT

Версия сервера zabbix 7.0.27. По LLD обнаруживаются названия процесса АВПО Касперский. Создаётся элемент данных, считывающий состояние процесса и триггер на остановку работы процесса. В случае остановки триггер должен вызвать действие старта процесса system.run[sc start <название процесcа АВПО>] Не могу понять как передать созданный обнаружением макрос названия АВПО в действие триггера.

Template for MariaDB Galera Cluster

studero — Mon, 27 Jul 2026 11:27:27 GMT

Hi,
I have create a MariaDB Galera Cluster, 3 nodes, on my infrastructure. Now, I want to monitor it with Zabbix. But I do not found a good template ? Only this template GitHub - MogiePete/zabbix-galera-template: Zabbix Template for Monitoring Galera in a linux environment. · GitHub but it is very old (9 years)!
I need help
Regards
Olivier

HTTP Agent causes hanging CLOSE_WAIT connections in Docker container

Shadown — Sun, 26 Jul 2026 18:44:26 GMT

Hello everyone,

I’m running into an issue with my Zabbix setup while monitoring a Proxmox environment using the official “Proxmox VE by HTTP” template.

Zabbix Server is running inside a Docker container.
Proxmox hosts are monitored via the official HTTP template.
The connection to the Proxmox API is made over port 8006.
The Proxmox host is reachable via IP, for example 192.168.210.2:8006.
Server version: 7.4.12

After a short runtime of about 15 to 30 minutes, the TCP connections to the Proxmox API get stuck in the CLOSE_WAIT state.

Code:

tcp        1      0 34a1a00fcd10:52880      192.168.210.2:8006      CLOSE_WAIT  
tcp        1      0 34a1a00fcd10:35162      192.168.210.2:8006      CLOSE_WAIT  
tcp        1      0 34a1a00fcd10:47348      192.168.210.2:8006      CLOSE_WAIT  
tcp        1      0 34a1a00fcd10:57528      192.168.210.2:8006      CLOSE_WAIT  
tcp        1      0 34a1a00fcd10:56086      192.168.210.2:8006      CLOSE_WAIT  
.
.
.

This is my Docker Compose configuration for the Zabbix server:

Code:

zabbix-server:
    image: zabbix/zabbix-server-pgsql:alpine-7.4-latest
    #ulimits:
    # nproc: 65535
    # nofile:
    #  soft: 20000
    #  hard: 40000
    container_name: zabbix-server
    restart: unless-stopped
    depends_on:
      - postgresql-server
    environment:
      DB_SERVER_HOST: postgresql-server
      POSTGRES_USER: ${POSTGRES_USER}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
      POSTGRES_DB: ${POSTGRES_DB}
      ZBX_NODEADDRESS: zabbix-server
      ZBX_NODEADDRESSPORT: 10051
      ZBX_STARTREPORTWRITERS: 3
      ZBX_WEBSERVICEURL: "http://zabbix-web-service:10053/report"
      ZBX_STARTHTTPPOLLERS: 1
      ZBX_DEBUGLEVEL: 3
      CURL_IPRESOLVE: 4
    ports:
      - "10051:10051"
    volumes:
      - zabbix-server-data:/var/lib/zabbix
      - zabbix-snmptraps-data:/var/lib/zabbix/snmptraps
      - zabbix-export-data:/var/lib/zabbix/export

When I enable the ulimit parameters, the issue still persists, but it takes longer to occur.