Zabbix internal data collection

Essentially, Zabbix is also an application system like other monitored systems. Therefore, its status and performance data also need to be collected and monitored. In order to understand its running performance and status, and then optimize its performance or perform necessary system hardware upgrades when necessary, or even change its architecture. For example, in an organization, the number of monitored devices and monitored projects increases rapidly as the business of the Organization expands and the actual development needs. After this growth reaches a certain level, the original Zabbix system of a single independent server model is difficult to meet the needs of the Organization, at this time, we may need to change the running mode of a single independent server to the distributed mode in a timely manner.

In fact, when we decide whether to upgrade the existing Zabbix system hardware, optimize the system performance, or even change the independent server mode to the distributed server mode, we need to face the basis. That is, when we decide to perform the above upgrade, we need data basis.

The internal data collection method of Zabbix system is to collect and monitor the status and performance data of Zabbix system. The internal data collection method of Zabbix is obtained by the Zabbix server through computation. Therefore, this data collection method does not need to be installed on any client. The following table 2-4 lists the monitoring items supported by the internal data collection method.

Table 2-4 List of internal data collection and monitoring items in Zabbix

Key)	Description	Remarks
Zabbix [boottime]	This metric item returns the Zabbix server-side process start time.	The return time is the number of seconds since, January 1, January 01, 1970. Therefore, you must manually convert the format to year, month, and day.
Zabbix [history]	Returns the number of records in the HISTORY Table in the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [history_log]	Return the data records in the HISTORY_LOG table of the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [history_str]	Return the data records in the HISTORY_STR table of the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [history_text]	Return the data records in the HISTORY_TEXT table of the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [history_uint]	Return the data records in the HISTORY_UINT table of the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as far as possible, so as not to have a greater impact on the database performance.
Zabbix [host, <type>, available]	Returns whether the specified host supports certain data collection methods. The value collected by this monitoring item corresponds to the data collection method possibility icon in the host list.	The type parameters in this project include agent, snmp, ipmi, and jmx. If the returned value is 0, the data collection method of the host is not used. If the returned value is 2, the availability of the collection method is unknown.
Zabbix [items]	Returns the total number of metric items that have been configured in the system.
Zabbix [items_unsupported]	Returns the total number of metric items configured in the database but not supported by the system.
Zabbix [java, <param>]	Returns information related to the Zabbix Java application gateway.	If <param> is ping, 1 is returned when the Java application gateway is available. For this metric item, you can use the nodata () function in the trigger to identify that the Java application gateway is unavailable. If the <param> parameter is version, the version information of the Java application gateway, such as 2.0.0, is returned.
Zabbix [process, <type >,< mode >,< state>]	This project returns the specified Zabbix process or a specified group of processes (specified by the <type> and <mode> parameters) the percentage of time in the <state> parameter. The returned value is calculated based on the data in the last minute. There are more than 20 Zabbix server-side processes (many zabbix system processes are often seen when we use the ps-ef command in the operating system, these system processes are called instances in zabbix. Different types of processes are formed when these instances are responsible for different jobs ). In this monitoring project, the <mode> parameter can be specified as a certain type of process number. Note that the process number here is not the process number in the system, but the sequence number of some type of process in the zabbix system. For example, if we use the StartPollers parameter in the configuration file to start five polling processes, the mode parameter can be set to 1 ~ 5. If the specified data exceeds this range, the created monitoring project reports an unsupported error. The maximum and minimum values here are for a single process. For example, we need to count the maximum or minimum values of a group of three types. If the "busy" rate of these three processes during the statistical period is 2, 18, and 66, the minimum "busy" rate of these processes is 2, the maximum "busy" rate is 66.	This metric item supports the following process types: Alarm (alerter)-this type of process is used to send alarm notifications; Configuration synx (configuration syncer) -- used to synchronize configuration information in the configuration file to the cache in the memory; Data sender-the process used by the server proxy node to send data (the server does not have such processes ); Database watchdog-this process is used to monitor the database status of the zabbix system. When the database status changes to unavailable, it sends a warning message (the server proxy does not support this type of process ). The automatic discoverer (discoverer) is used to automatically discover the process of the device; STEP (escalator)-the process used to process the steps in the action; Heartbeat sender-the server proxy sends heartbeat information (the server does not have this type of process ); History syncer: used to write historical data tables; Housekeeper: a process used to clear expired historical data; HTTP poller: Used to poll monitoring items of the web class; Ping Checker (icmp pinger) -- used for regular icmp ping checks; Ipmi Poler-used to regularly check ipmi monitoring items; Java poeller-used to poll java monitoring projects; Distributed node watcher is a process used to send historical data and update configuration information on different distributed nodes; Poller: used for polling of common passive monitoring items; Proxy poller: used for passive polling of the server proxy; Self-monitoring (self-monitoring)-used to collect internal monitoring information of Zabbix; Timer (timer)-used to process time-related functions and maintenance mode in the trigger; Trapper (trapper): used to process active collection, synchronization, and communication between distributed nodes or server proxies; Unreachable poller: Used to poll devices that cannot be reached; Vmware collector-collects data from vmware service processes (server agents do not support such processes ); Available mode parameters include: Avg -- average value of all processes of the specified type; Count -- returns the number of Created processes of the specified type; Max -- maximum value; Min -- Minimum value; <Process number> -- process number. For the meaning, see the description in "Description; Available state parameters include: Busy: indicates a busy process; Idle: A process in idle state;
Zabbix [proxy, <name >,< param>]	This metric is used to collect information about Zabbix server proxy.	The <name> parameter is the name of the server proxy. Available <param> is lastaccess: indicates the last time the server proxy information was received. The trigger function fuzzytime () can be used to check the availability of the server proxy.
Zabbix [proxy_history]	Return the number of data records that are waiting to be sent to the server in the proxy history Table on the server.
Zabbix [queue, <from>, <to>]	Returns the number of metric items that are <from> to <to> delayed before data is collected in the system queue.	<From> the default value is 6 seconds; <To> the default value is infinity;
Zabbix [rcache, <cache>, <mode>]	Return the memory space usage information used to cache configuration information in the memory.	Optional buffer parameter cache; You can select the mode parameter: Total -- returns the total memory size used to cache configuration information. The returned value must be the same as the value specified by the CacheSize configuration option in the configuration file; Free -- return the idle memory size used to cache configuration information; Pfree -- returns the percentage of idle memory space used to cache configuration information to the total memory space; Used -- return the memory space used to cache configuration information;
Zabbix [requiredperformance]	This metric item returns the expected performance value of the Zabbix server, that is, the number and type of monitored hosts and monitored items configured in the Zabbix system, the estimated number of new data to be processed by the server per second.	The returned value of this metric item is in the "status statistics"-> "status panel"-> "ZABBIX status" and "system report"-> "ZABBIX status ""server performance (value/second) "The project displays the same value and has the same meaning.
Zabbix [trends]	Return the record data in the TRENDS table in the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [trends_uint]	Return the data records in the TRENDS_UINT table of the database.	If the database is the storage engine of MySQL InnoDB or Oracle or PostgreSQL, do not use this metric item as much as possible, so as to avoid a great impact on the database performance.
Zabbix [triggers]	Returns the number of triggers configured in the database.
Zabbix [wcache, <cache >,< mode>]	This metric item is used to collect write Cache Information in Zabbix system.	You can select values, history, trend, and text for parameter cache. When the values item is selected for the cache parameter, the mode parameter can be set to the following values: All -- returns the total number of metric data processed since the zabbix server process or server proxy process is started. However, this does not include unsupported monitoring items. Float -- returns the sum of all floating point monitoring data processed since the zabbix server process or server proxy process is started. Uint -- returns the sum of all unsigned integer metric data processed since the zabbix server process or server proxy process is started. Str -- returns the total number of string-type metric data processed since the zabbix server process or server proxy process is started. Log -- returns the total number of log-type metric data processed since the zabbix server process or server proxy process is started. Text -- returns the sum of all text-type metric data processed since the zabbix server process or server proxy process is started. Not supported -- return the total number of unsupported monitoring data processed by the system since the zabbix server process or the server proxy process is started. When the history item is selected for the cache parameter, the mode parameter can be set to the following values: Pfree -- returns the percentage of idle buffers in the History buffer zone in the total history buffer zone. The total size of the historical buffer is specified by the HistoryCacheSize parameter in the configuration file. The historical buffer zone is used to cache all types of monitoring project information, the time of data collection, and the data collected by numerical monitoring projects. If the value of this project is low, it is likely that there is a performance problem in the database. Free -- returns the size of the idle historical buffer. Total -- returns the total size of the historical buffer. Used -- returns the size of the used historical buffer. When the trend item is selected for the cache parameter, the mode parameter can be set to the following values: Pfree -- returns the percentage of idle buffers in the trend buffer to the total trend buffer. The total size of the trend buffer is specified by the TrendCacheSize parameter in the configuration file. The trend buffer is used to cache all data collected by the monitoring project in the last hour for data aggregation. The server proxy does not support this data collection method. Other statistical items in the trend Buffer: free, total, and used are similar to those in the historical buffer. Here we will not describe them one by one. When the text item is selected for the cache parameter, the mode parameter can be set to the following values: Pfree -- returns the percentage of free buffers in the text buffer to the total text buffer. The total text buffer size is specified by the HistoryTextCacheSize parameter in the configuration file. The text buffer is used to cache historical data of characters and text in log type. However, the information of these types of monitoring items and the time cut-off information of the collected data are still stored in the historical buffer zone. Other statistical items in the text Buffer: free, total, and used are similar to those in the History buffer. Here we will not describe them one by one.