The expressions used in triggers are very flexible. You can use them to create complex logical tests regarding monitored statistics.
A simple useful expression might look like:
Trigger functions allow to reference the collected values, current time and other factors.
A complete list of supported functions is available.
Most of numeric functions accept the number of seconds as a parameter.
You may use the prefix # to specify that a parameter has a different meaning:
FUNCTION CALL | MEANING |
---|---|
sum(600) | Sum of all values within 600 seconds |
sum(#5) | Sum of the last 5 values |
The function last uses a different meaning for values when prefixed with the hash mark - it makes it choose the n-th previous value, so given the values 3, 7, 2, 6, 5 (from most recent to least recent), last(#2) would return 7 and last(#5) would return 5.
A parameter must be given even for those functions which ignore it. Example: last(0)
Several functions support an additional, second time_shift
parameter. This parameter allows to reference data from a period of time in the past. For example, avg(1h,1d) will return the average value for an hour one day ago.
Triggers only evaluate history information. If history is not available (especially relevant for time shift), trend information is not used, thus history must be kept for at least the period trigger functions expect it.
You can use the supported unit symbols in trigger expressions, for example '5m' (minutes) instead of '300' seconds or '1d' (day) instead of '86400' seconds. '1K' will stand for '1024' bytes.
The following operators are supported for triggers (in descending priority of execution):
PRIORITY | OPERATOR | DEFINITION | |
---|---|---|---|
1 | / | Division | |
2 | *** |Multiplication | |3** | - | Arithmetical minus |
4 | + | Arithmetical plus | |
5 | < | Less than. The operator is defined as: A<B ⇔ (A<=B-0.000001) |
|
6 | > | More than. The operator is defined as: A>B ⇔ (A>=B+0.000001) |
|
7 | # | Not equal. The operator is defined as: A#B ⇔ (A<=B-0.000001) | (A>=B+0.000001) |
|
8 | = | Is equal. The operator is defined as: A=B ⇔ (A>B-0.000001) & (A<B+0.000001) |
|
9 | & | Logical AND | |
10 | | | Logical OR |
Processor load is too high on www.zabbix.com
'www.zabbix.com:system.cpu.load[all,avg1]' gives a short name of the monitored parameter. It specifies that the server is 'www.zabbix.com' and the key being monitored is 'system.cpu.load[all,avg1]'. By using the function 'last()', we are referring to the most recent value. Finally, '>5' means that the trigger is in the PROBLEM state whenever the most recent processor load measurement from www.zabbix.com is greater than 5.
www.zabbix.com is overloaded
{www.zabbix.com:system.cpu.load[all,avg1].last(0)}>5|{www.zabbix.com:system.cpu.load[all,avg1].min(10m)}>2
The expression is true when either the current processor load is more than 5 or the processor load was more than 2 during last 10 minutes.
/etc/passwd has been changed
Use of function diff:
The expression is true when the previous value of checksum of /etc/passwd differs from the most recent one.
Similar expressions could be useful to monitor changes in important files, such as /etc/passwd, /etc/inetd.conf, /kernel, etc.
Someone is downloading a large file from the Internet
Use of function min:
The expression is true when number of received bytes on eth0 is more than 100 KB within last 5 minutes.
Both nodes of clustered SMTP server are down
Note use of two different hosts in one expression:
{smtp1.zabbix.com:net.tcp.service[smtp].last(0)}=0&{smtp2.zabbix.com:net.tcp.service[smtp].last(0)}=0
The expression is true when both SMTP servers are down on both smtp1.zabbix.com and smtp2.zabbix.com.
Zabbix agent needs to be upgraded
Use of function str():
The expression is true if Zabbix agent has version beta8 (presumably 1.0beta8).
Server is unreachable
The expression is true if host "zabbix.zabbix.com" is unreachable more than 5 times in the last 30 minutes.
No heartbeats within last 3 minutes
Use of function nodata():
To make use of this trigger, 'tick' must be defined as a Zabbix trapper item. The host should periodically send data for this item using zabbix_sender. If no data is received within 180 seconds, the trigger value becomes PROBLEM.
Note that 'nodata' can be used for any item type.
CPU activity at night time
Use of function time():
{zabbix:system.cpu.load[all,avg1].min(5m)}>2&{zabbix:system.cpu.load[all,avg1].time(0)}>000000&{zabbix:system.cpu.load[all,avg1].time(0)}<060000
The trigger may change its status to true, only at night (00:00-06:00) time.
Check if client local time is in sync with Zabbix server time
Use of function fuzzytime():
The trigger will change to the problem state in case when local time on server MySQL_DB and Zabbix server differs by more than 10 seconds.
Comparing average load today with average load of the same time yesterday (using a second time_shift
parameter).
This expression will fire if the average load for the last hour tops the average load of the same hour yesterday more than two times.
Sometimes a trigger must have different conditions for different states. For example, we would like to define a trigger which would become PROBLEM when server room temperature is higher than 20C while it should stay in the state until temperature will not become lower than 15C.
In order to do this, we define the following trigger:
Temperature in server room is too high
Note the use of a macro {TRIGGER.VALUE}. The macro returns current trigger value.
Free disk space is too low
Problem: it is less than 10GB for last 5 minutes
Recovery: it is more than 40GB for last 10 minutes
({TRIGGER.VALUE}=0&{server:vfs.fs.size[/,free].max(5m)}<10G) |
({TRIGGER.VALUE}=1&{server:vfs.fs.size[/,free].min(10m)}<40G)
Note use of macro {TRIGGER.VALUE}. The macro returns current trigger value.