Previously, working in some other location than Status of triggers or Dashboard pages would not show any information regarding issues that are currently happening. Starting with 1.8.3, support for global frontend notifications is added. Global notifications involve both showing a message and playing a sound.
Global notifications can be enabled per user in profile configuration. If enabled, global message timeout can be changed. By default, messages will stay on screen for 90 seconds.
Additionally, it is possible to configure details of sound notifications. There are three possible options for playing a sound:
It is possible to receive messages for problems and for resolutions. Messages can be filtered based on trigger severity as well. For each trigger severity and recovery message sound to be played can be customised.
As the messages arrive, they are displayed in a floating section on the right hand side. This section can be repositioned vertically.
For this section, several controls are available:
A new frontend theme has been added - Dark orange.
A new per-map option has been introduced - ability to mark map elements that have their state changed recently.
When this option is enabled, any map elements that recently had state changed (trigger firing or resolving) are marked by red triangles.
These triangles are placed on top, bottom, left and right of an object. If the object has a label, no triangle is displayed at the label location.
The period of showing these marks is the same as for trigger state flashing - 30 minutes.
In previous 1.8 branch versions of Zabbix it was not easy enough to place icons in exact locations and align them nicely - with the gain of drag and drop ability to enter element position manually was lost. Now this ability is restored.
1.8.2 and before | 1.8.3 |
An even bigger improvement - map icon aligning was added.
By default, grid for aligning is 50x50. Available grids:
If aligning is on, moving an icon will align it at the center of the grid cell. Clicking button Align icons will align all icons to the nearest grid cell. It is now also possible to hide grid while editing the map.
It is now easier to edit status indicators for map links - line style and colour can be changed directly from the list:
Additionally, it is now possible to add multiple indicators with same line style and colour in one go:
For the previous form, multiple triggers can be selected with checkboxes in the trigger list.
Available as a part of network map export and import, image export and import has been implemented. In the map import dialogue, there are checkboxes for icon or map background importing. This option is only available to Zabbix Super Administrators.
1.8.2 and before | 1.8.3 | |
Disabled | ||
In maintenance |
It is now possible to customise dashboard display depending on trigger acknowledgement state.
In the dashboard filter options, there is a new dropdown available, Problem display.
It provides 3 options:
The difference between these 3 options can be seen here:
All | Separated | Unacknowledged only |
When in the Separated mode, tooltip for each number shows the corresponding issues - either the unacknowledged ones, or all of them.
Configuration section in 1.8.3 has been changed to improve usability.
Trigger line in Zabbix graphs was extremely hard to spot in latest Zabbix versions. In 1.8.3, the line has been changed from 1 pixel height to 2 pixels. It also adds two additional changes:
These changes should change the trigger line from nearly invisible to very clearly visible - let's compare the 1.8.2 version and 1.8.3 one. In 1.8.2 trigger line is invisible unless the observer knows it's there.
Trigger line in 1.8.2 | Trigger line in 1.8.3 |
Log viewing filter was brought in line with other filters in regard to being hideable. It was also redesigned, especially to make handling of multiple logfiles easier.
Log filter in 1.8.2 and before | Log filter in 1.8.3 |
Additionally, this filter is also available now for text items, not for log items only.
Zabbix 1.8.3 has received two improvements to the global search functionality.
Global search now provides suggestions, based on the entered string. This string is matched against hosts only and all suggestions are lowercase independent on the actual case of the hostname.
Search results form has been greatly improved.
While the new search results form is wider, it should be much more functional now:
In this screenshot, user has write access only to host First Linux Server, so editing options for other hosts are not available.
By default, specifying a unit for an item results in the multiplier prefix being added - for example, value '2048' with unit B would be displayed as 2KB. For a pre-defined, hard-coded list of units this is now prevented:
Support for the following units has been added for graph y axis:
Note that smallest available units might limited by available database field precision.
It is now possible to choose All as a period in graphs and elsewhere.
Maximal period currently is limited to two years.
Previously, changing any parameter in the graph configuration would regenerate the graph. Now it only happens when really necessary, and button Preview has been added to force regenerating.
Macro {TRIGGER.EVENTS.UNACK} now is supported in map element labels. Additionally, these macros now expand to the following values:
Macro | Value |
---|---|
{TRIGGER.EVENTS.UNACK} | Number of unacknowledged events for a map element in maps, or for the trigger which generated current event in notifications. |
{TRIGGER.EVENTS.PROBLEM.UNACK} | Number of unacknowledged PROBLEM events for all triggers disregarding their state. |
{TRIGGERS.UNACK} | Number of unacknowledged triggers for a map element, disregarding trigger state. |
{TRIGGERS.PROBLEM.UNACK} | Number of unacknowledged PROBLEM triggers for a map element. |
{TRIGGER.PROBLEM.EVENTS.PROBLEM.UNACK} | Number of unacknowledged PROBLEM events for triggers in PROBLEM state. |
The following new macros have been added:
Macro | Value |
---|---|
{TRIGGER.EVENTS.ACK} | Number of acknowledged events for a map element in maps, or for the trigger which generated current event in notifications. |
{TRIGGER.EVENTS.PROBLEM.ACK} | Number of acknowledged PROBLEM events for all triggers disregarding their state. |
{TRIGGER.PROBLEM.EVENTS.PROBLEM.ACK} | Number of acknowledged PROBLEM events for triggers in PROBLEM state. |
{TRIGGERS.ACK} | Number of acknowledged triggers for a map element, disregarding trigger state. |
{TRIGGERS.PROBLEM.ACK} | Number of acknowledged PROBLEM triggers for a map element. |
Proxy view in Administration → DM → Proxies has been improved by adding two new columns - item count and required performance (showing same information as in the server status report, new values per second). That should help with proxy hardware requirement estimates.
If an item or a trigger is coming from a nested template chain, it isn't always obvious how exactly is it attached to the host. Since 1.8.3, in item or trigger editing form header full inheritance chain is shown, and it is possible to click on any template to see item editing form for that template.
|<| |<| |-|
|<| |<| |-|
In 1.8.2 and before, text item history had lots of vertical whitespace. Because of this less lines could fit in a screen. In 1.8.3, vertical whitespace has been reduced considerably.
Text data in 1.8.2 and before | Text data in 1.8.3 |
Note that in Zabbix history viewing newer items are at the top, so you have to read lines from bottom up.
While logfile viewing did not have so much excessive vertical whitespace, it could still be slightly reduced even when adding table lines in 1.8.3.
Log data in 1.8.2 and before | Log data in 1.8.3 |
In previous versions, y axis labels for Zabbix graphs were displayed at even distances for full graph scale. In some cases, this resulted in label values that were hard to read. For example, graph going from 0 to 100 would have labels including 18.33, 36.67 and 91.67. In Zabbix 1.8.3, label calculation is improved. Zabbix will split the scale based on some round value, thus in our example values would become 20, 40 and 80. Additionally, if the graph has manual maximum scale entered, it is always printed. If printing the maximum would collide with last rounded value, last rounded value is hidden. For example, on scale from 0 to 110 both of the labels 100 and 110 are printed. On scale from 0 to 101 only the label 101 is printed.
These benefits are visible both with large and small graph values.
Large values | Small values | |||
---|---|---|---|---|
1.8.2 and before | 1.8.3 | 1.8.2 and before | 1.8.3 |
It is now possible to "unlink and clear" (which not only unlinks the templates, but also deletes the linked elements) templates for host mass update. For this purpose, checkbox "Clear when unlinking" must be marked in the "Replace linked templates" section of the host mass update screen:
The filter in the "Status of Triggers" page has been improved. Trigger acknowledge status has been brought out in a separate menu, and a new filtering option is available - "With last event unacknowledged". With these changes it is again possible to display all (including those in OK state) triggers for all hosts and host groups.
It is also now possible to filter triggers by age - option "Age less than" allows to only show those triggers that have changed state in last N days.
Additionally, several options have been renamed to make it more clear what they are about - for example, Select became Filter by name.
Instead of providing fixed periods, slideshow refresh can now be set to multiples of default settings. This allows to slow down or speed up slideshow displaying while still maintaining relative time of displaying for each slide.
The following new translations have been added:
The following translations have been updated:
Highlighting in 1.8.2 and before | |
Highlighting in 1.8.3 |
Up to Zabbix version 1.8.3 Zabbix proxy was the one connecting to Zabbix server, and that was the only supported mode. But if one would like to place Zabbix server inside the local network which would monitor host in DMZ by proxy, that would not be possible - connection would have to go the other way.
Starting with version 1.8.3, Zabbix server and proxy communication can be in either direction. To control this, several new options have been introduced.
In addition to previously supported zabbix[history] and zabbix[trends] items, Zabbix now supports additional items to monitor amount of values in corresponding tables. Note that on most database engines usage of these items can seriously degrade the performance.
The following simple checks have been added:
Zabbix internal item zabbix[queue] gained parameter support. It is now possible to specify for how long item must be missing data to be counted. For example, zabbix[queue,6,59] will count all items that are late by 6-59 seconds, inclusive. Time based suffixes are supported for these parameters, so the following syntax will check for all items that are haven't been refreshed between one minute and 6 hours: zabbix[queue,1m,6h].
By default first value is 6, and second value is empty, which means infinity.
This version changes how acknowledge works. In previous versions, if a trigger was acknowledged and changed state to UNKNOWN, then back, this new event would not be acknowledged (and in turn, trigger would become un-acknowledged). Since version 1.8.3, any acknowledge is copied to all events of the same type that are separated by UNKNOWN events. For example, if events have happened like this (newer events on top):
Acknowledging event number 3 in the frontend would add the same acknowledge to event 5, and vice versa. If new events would now happen:
In this case Zabbix server would add the acknowledge (if any) from event 9 to the new PROBLEM event number 11.
There is one exception - multiple TRUE (PROBLEM) triggers do not get the acknowledges copied over from other events.
Upcoming MySQL version 5.5 drops support for table keyword type, and using ENGINE is required. Zabbix 1.8.3 changes to usage of keyword ENGINE.