Docs

The Scatter plot widget has been added to dashboard widgets. It displays the relationship between two metrics by plotting individual data points along an X and Y axis. This helps reveal patterns, clusters, correlations, and outliers in the data set.

Graphs in the Graph widget can now be displayed with inverted Y-axis values. A new Invert values setting allows you to multiply Y-axis values by −1 without altering the original data.

Templates

New templates

Aruba CX 8300s by SNMP, a template providing SNMP-based monitoring for the Aruba CX 8300 switch series.
The AWS by HTTP template set has been supplemented with the template AWS Backup Vault by HTTP.
The Azure by HTTP template set has been supplemented with the template Azure Sentinel by HTTP.
Ciena 3906 by SNMP, a template for monitoring Ciena 3906 devices.
Cisco Secure Firewall Threat Defense by HTTP, a template providing monitoring capabilities for Cisco Secure Firewall Threat Defense devices using REST API.
Cradlepoint NCM v2 by HTTP, a template set for monitoring Cradlepoint NCM v2 and its devices via HTTP.
Domain RDAP by HTTP, a template for monitoring domain registration data via the RDAP protocol.
IBM Maximo Service Request, a webhook-based template that enables linking Zabbix to IBM Maximo.
MariaDB by ODBC, a template for monitoring MariaDB databases via ODBC.
Microsoft Hyper-V Failover Cluster by SSH and Microsoft Hyper-V Standalone by SSH, templates for monitoring Microsoft Hyper-V clusters and standalone hosts via SSH.
OpenAI Platform by HTTP, a template for monitoring OpenAI's developer platform.
Percona by ODBC, a template for monitoring Percona databases via ODBC.
Ribbon SBC Edge by HTTP, a template for monitoring Ribbon SBC Edge (formerly SWe Lite) devices via HTTP.
Ribbon SBC SWe Core by HTTP, a template for monitoring SBC SWe Core devices via HTTP, and Ribbon SBC SWe CE by HTTP, a template for monitoring Ribbon SBC SWe Call Engine (CE) instances via HTTP.
Stormshield SNS by SNMP, a template for monitoring Stormshield Network Security (SNS) devices via SNMP.
VeloCloud SD-WAN Edge by HTTP, a template for monitoring VeloCloud SD-WAN Edge devices via HTTP.
Vyatta Virtual Router by SNMP, a template for monitoring the Vyatta 1908e virtual router.

Updated templates

Ciena 3906 by SNMP has been updated to include filesystem and CPU load items.
The GCP by HTTP template set has been supplemented with the template GCP Cloud Run Service by HTTP.
GitHub organization by HTTP has been updated to include items for monitoring Microsoft Copilot.
The GLPi webhook media type has been updated to support the RESTful API v2 with OAuth2 authorization (enabled by default), optional legacy REST API v1 support via configuration flag, ticket creation for problem events, and user-configurable urgency mapping.
Microsoft 365 reports by HTTP has been updated to include items for monitoring Microsoft Copilot.
MySQL by Zabbix agent, MySQL by Zabbix agent 2, MySQL by Zabbix agent active, MySQL by Zabbix agent 2 active, and MySQL by ODBC have been updated to support both SHOW SLAVE STATUS (old syntax) and SHOW REPLICA STATUS (new syntax).
MySQL by ODBC has also been updated with new metrics, discovery rules for tables and replicas, and an improved dashboard.
PostgreSQL by ODBC has been updated with new metrics, discovery rules, and dashboards, along with general improvements to performance and observability.
Proxmox VE by HTTP has been updated with the nested LLD functionality. In addition, the unit format for items displaying percentages has been adjusted for clarity.
VeloCloud SD-WAN by HTTP, previously VMWare SD-WAN VeloCloud by HTTP, has been renamed and updated to align with the latest VeloCloud SD-WAN platform.
Vyatta Virtual Router by SNMP has been updated with new OID items and dashboard enhancements.

Items

JSON data type

Zabbix now supports JSON as a data type for item values.

Previously, JSON values were collected by text items and stored as strings with a limit of 64KB. Now, Zabbix can store JSON values natively with a limit of 128MiB, and also reject invalid JSON values (e.g., containing unquoted keys, trailing commas, mismatched brackets).

The JSON data type is supported by all item types and item prototypes (except Calculated), and is available in real-time data export and connectors. JSON values can be stored in all supported databases and Elasticsearch. If you're using TimescaleDB, please see upgrade notes.

Note that JSON items cannot be used in triggers; however, you can extract JSON fields with dependent items that have a non-JSON data type and use them in triggers.

Items that return a JSON string (net.if.discovery, vfs.file.get, etc.) are still text items; however, you can change them to JSON if needed.

For more details and JSON data limits, see item configuration.

Type parameter for S.M.A.R.T disk discovery

The smart.disk.discovery item (Zabbix agent 2 S.M.A.R.T. plugin) now accepts an optional type parameter to specify a value to scan for the disks.

Plugins

Ceph plugin

This plugin now operates in two modes:

native - This mode uses the go-ceph library to communicate directly with the Ceph cluster using the native Ceph API (msgr2 protocol). This is the recommended mode for modern Ceph installations, but is only supported on Linux and starting with Ceph 16.
restful (deprecated) - This mode uses the Ceph RESTful API for communication. It is the default mode for backward compatibility but will not work with Ceph version 20 (Tentacle) or newer due to the removal of the mgr/restful module.

Which mode to use is determined by the value of the mode parameter (native/restful):

Plugins.Ceph.Default.Mode=native - set the native mode for the plugin
Plugins.Ceph.Sessions.<SessionName>.Mode=native - set the native mode for the named session

Note that the user credential set differs for each mode and they are not compatible with each other. The Plugins.Ceph.InsecureSkipVerify parameter is ignored in native mode because connection security is defined on the Ceph cluster side by msgr2 protocol (secure by default).

Note that the Ceph plugin for Zabbix agent 2 is now a loadable plugin and requires additional installation steps. This is due to the librados package requirement (for the native mode). See the Ceph plugin readme for details.

MongoDB plugin

Zabbix agent 2 MongoDB plugin offers multiple enhancements:

Full support for conventional MongoDB URI parsing, with support for both mongodb:// and mongodb+srv:// schemes.
Automatic discovery of MongoDB replica set nodes, enhancing the ability to monitor MongoDB clusters.
Support for x509 authentication, enabling secure connections to MongoDB using client certificates.

Oracle plugin

Zabbix agent 2 Oracle plugin now supports encrypted connections to Oracle databases using the TCPS (TLS) protocol. This adds the ability to monitor Oracle instances over TLS-secured sockets, improving security for remote monitoring.

Redis plugin — TLS support and startup-time validation

TLS support has been added to the Redis plugin for Zabbix agent 2.

Startup-time validation of the plugin TLS configuration was implemented and validation/error messages improved. An invalid configuration logic (for example: using connection type verify_full without specifying TLSCAFile) can prevent Zabbix agent 2 from starting.

Test execution mode for loadable plugins

Loadable plugins can now be launched in test mode using the -t (--test) flag, passing an item key as argument. In this mode, the plugin runs for debugging and development purposes, and plugin configuration files are ignored.

Low-level discovery

Convert to JSON checkbox in discovery forms

A new Convert to JSON checkbox was added to the low-level discovery rule form and discovery prototype form, which is shown if "HTTP agent" is selected in the Type dropdown. This option allows to mark retrieved data for automatic conversion to JSON before further processing.

Macro support for nested low-level discovery

Low-level discovery macros are now supported in nested low-level discovery rules, in:

JSONPath preprocessing parameters
JSONPath field for custom LLD macros

Editable tags for triggers created by low-level discovery

Triggers created from trigger prototypes can now have tags added manually. Tags inherited from trigger prototypes are still applied automatically. Tags added manually may be modified on discovered triggers and will be included in event tag arrays and available to tag-aware features such as filters, dashboard widgets and notification macros.

Processes

SNMPv3 EngineID caching and reuse

Zabbix now caches SNMPv3 EngineID → IP mappings and will attempt to reuse cached EngineIDs for subsequent SNMPv3 checks, reducing probe traffic and improving poller performance. If a reused EngineID does not respond, the poller falls back to an EngineID probe and may remove stale entries after interface changes or persistent failures.

Refined proxy throttling during history cache recovery

Proxy throttling logic has been refined to improve server stability during history cache recovery. When history cache usage reaches the throttling threshold, the server continues to stop accepting proxy data as before. When cache usage falls to 60%, the server begins processing the throttling list but may still reject proxy uploads that contain very large batches (approximately more than 10k records) until cache pressure decreases further. This change reduces the risk of repeated cache overloads while the server recovers.

Increased maximum timeout for the zabbix_get and zabbix_js

The maximum value for the timeout parameter of the zabbix_get and zabbix_js command-line utilities has been increased to 600 seconds.

Optimized housekeeping

Housekeeping has been optimized by moving the creation of housekeeper tasks to database triggers. When items (including low-level discovery rules), triggers, services, or network discovery rules are deleted, database triggers (instead of explicit application-level calls) now populate the housekeeper table with cleanup tasks. For details, see the housekeeping procedure.

The housekeeper now also removes network discovery events generated from discovered hosts or services that have since been deleted, as well as all events associated with problems of triggers that have since been deleted (previously, only the problems themselves were removed; associated events were only removed after the housekeeper Trigger data storage period expired).

Manual close — trigger tags inherited by recovery events

Recovery events created after a manual close inherit trigger tags in addition to item and host tags. These tags are present in the event tags array and are available to notification macros such as {EVENT.RECOVERY.TAGS} and {EVENT.RECOVERY.TAGSJSON}.

DNS query caching for Zabbix daemons

Zabbix server, Zabbix proxy and Zabbix agent received support for using the c-ares resolver for all DNS requests, providing DNS query caching and improved resolver failover when Zabbix is built with --with-ares. DNS query caching requires c-ares 1.26.0 or later.

Build support for c-ares on Windows

Zabbix agent can now be built with the c-ares resolver on Microsoft Windows. c-ares may be installed via vcpkg, and the agent build supports either ARES=<vcpkg prefix> or separate ARESINCDIR/ARESLIBDIR paths for include and library directories.

Authentication

Importable SAML certificates for single sign-on

Super admin users can now import certificates and private keys directly in the frontend for SAML configuration. Three new fields were added to Administration > Authentication > SAML:

IdP certificate - X.509 certificate presented by the identity provider
SP certificate - service provider certificate used for SAML exchanges
SP private key - private key corresponding to the SP certificate

These controls allow adding new values or modifying existing ones from the SAML settings tab. Certificates and private keys are validated before being saved to the chosen storage backend; invalid or malformed values are rejected with an explanatory error.

Widgets

Geomap host marker clustering

The Geomap widget now supports a Clustering parameter to control how nearby host markers are combined into a single marker with a count. By setting the map zoom level for clustering, you can keep large maps readable and maintain precise visibility when zoomed in.

Aggregate columns in Top items

The Top items widget can now group item patterns into a single aggregated column or row, making it easy to present combined metrics (for example: total incoming traffic across all network interfaces of a host). Three new options were added when configuring item columns:Aggregate columns, Column aggregation function, and Combined column name.

The Graph widget offers multiple usability enhancements:

Items in the graph data tooltip that appears when hovering over the graph are sorted by value in descending order.
Hovering over an item in the tooltip highlights its graph while dimming others.
Selecting an item in the tooltip broadcasts its data to other widgets that support listening to it.
The limit on the number of items displayed in the tooltip has been removed.
A new Host names in labels setting allows you to choose whether host names are displayed in the tooltip and graph legend.

A new Invert values setting allows you to flip a graph.

Frontend

New and embedded fonts

Zabbix now includes new and embedded fonts that improve readability, load faster, and render additional frontend languages with minimal layout impact. Being embedded, these fonts also ensure a consistent appearance across systems.


New fonts:
Legacy fonts:

The new fonts are used in almost all themes. If needed, the newly added Blue (classic) and Dark (classic) themes are available with legacy fonts.

The font family for monospace content and graphs remains unchanged.

Inline validation

The following forms in the frontend have been added to the set of forms that support inline validation:

Input errors are displayed immediately after filling in the fields, improving usability and reducing configuration mistakes.

In low-level discovery setup, the host prototype configuration form is now opened in a modal (pop-up) window.

Inherited tags visible in hosts, templates, items, and triggers

Inherited tags are now displayed and returned consistently across templates, hosts, items, web scenarios, and triggers. Tags inherited from template/host chains are visible in Monitoring > Latest data. Filtering by inherited tags is available wherever tag filtering is supported — including the Monitoring > Latest data and Data collection sections, and all dashboard widgets that allow filtering by host, item, trigger, or web scenario tags — so tag-based selection and subfiltering behave the same regardless of where a tag was defined.

The Tags tab on template, host and host prototype configuration forms now exposes a radio element to choose how tags are presented: templates show Template tags / Inherited and template tags, and hosts and host prototypes show Host tags / Inherited and host tags. Inherited tags are visually distinguished by a new outlined document icon next to the tag label.

Graph and Pie chart widgets include a new Item tags setting.

Tooltips positioning

Tooltips can now be repositioned by dragging them to a new location. This applies, for example, to the Graph tooltip, as well as to tooltips with descriptions in the Latest Data or Problems sections.

Documentation

Consolidated documentation pages for minor releases

Release documentation for minor versions of a major Zabbix release will now be collected in single documentation pages for new features and upgrade notes respectively.

What’s next?

What's new in Zabbix 8.0.x