Aggregate CPU metrics not updating in calculated items

[Jose Garcia]

Hi everyone,

I’ve encountered a problem with calculated items.

I’m using the following setup:

• Debian v12.11
• Zabbix server v6.0.41
• PostgreSQL v15.13
• TimescaleDB v2.15.2

I'm collecting CPU items using an LLD rule, defined as follows:

• Name: CPU discovery
• Key: cpu.discovery
• SNMP OID: discovery[{#SNMPVALUE},.1.3.6.1.4.1.9.9.109.1.1.1.1.2]

This correctly discovers all CPUs on the host and creates individual items, such as:

• Name: CPU utilization
• Key: cpu.util[{#SNMPINDEX}]
• SNMP OID: .1.3.6.1.4.1.9.9.109.1.1.1.1.8.{#SNMPINDEX}

For example, the resulting keys might be:

• cpu.util[1]
• cpu.util[2]
• cpu.util[3]
• cpu.util[4]

These metrics are collected every 5 minutes, with a preprocessing step: "Discard unchanged with heartbeat" set to 1 hour.

Then, I aggregate these metrics using a calculated item:

• Name: CPU utilization
• Key: cpu.util
• Formula: avg(last_foreach(//cpu.util))

This calculated item also runs every 5 minutes and has the same preprocessing: "Discard unchanged with heartbeat" up to 1 hour.

As far as I understand, this should calculate the average value across all CPU items.

However, the value remains the same for several days, even though the individual CPU metrics are being updated.

Has anyone experienced a similar issue?

Thank you for your help.