Modules

NEB

The NEB module is required if monitoring events generated by a monitoring engine (Centreon Engine, Nagios, Icinga, ...) are transmitted by this instance of Centreon Broker. No special configuration is required, the module just need to be loaded through the global module_directory directive.

File

As its name suggests, the file module is used to read from or write to files.

Type file
Layer(s) 1-3
Work on input Yes
Work on output Yes
Work on temporary Yes

Configuration

Tag Description
path Path to the file.
max_size Maximum file size in bytes. If the limit is reached, Broker will go on with path1, then path2, ..., pathN.

Example

<input>
  <type>file</type>
  <path>/var/lib/centreon/broker.file</path>
  <protocol>bbdo</protocol>
  <max_size>100000000</max_size> <!-- 100MB limit -->
</input>

Local

The local module allow local domain (a.k.a. Unix domain) socket creation. Sockets can be created either as clients (type local_client) or server (type local_server).

Type local_client or local_server
Layer(s) 1-3
Work on input Yes
Work on output Yes
Work on temporary No

Configuration

Tag Description
one_peer_retention_mode Allow only one connection for a specific output. This option allow to keep retention on a server output (socket on a listen mode).
path Path of the local socket.
protocol Choose your protocol (bbdo or ndo).

Example

<output>
  <type>local_client</type>
  <path>/var/lib/centreon/broker.sock</path>
  <protocol>bbdo</protocol>
</output>
<output>
  <type>local_server</type>
  <path>/var/lib/centreon/broker.sock</path>
  <protocol>bbdo</protocol>
</output>
<output>
  <type>local_server</type>
  <path>/var/lib/centreon/broker.sock</path>
  <protocol>bbdo</protocol>
  <one_peer_retention_mode>1</one_peer_retention_mode>
</output>

RRD

The RRD module is used to create RRD graphs. It fetches events generated by storage output and directly handle them. It was designed to be used with Centreon and replace the Centstorage software. Its type in the configuration should be rrd.

Type rrd
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Description

Notice that an RRD endpoint receives data to insert in RRD graphs from a storage endpoint. RRD endpoints only manage physical graph creation, update and deletion. The whole intelligence on what to update and when is not done here.

Configuration

Tag Description
metrics_path Path to where the metrics graphs should be written.
status_path Path to where the status graphs should be written.
path If using rrdcached software with local socket connection, path to this socket.
port If using rrdcached software with a network connection, port on which rrdcached listens. The RRD module onlys supports connection with localhost rrdcached.
write_metrics Enable or disable metrics graph creation and update. Enabled by default.
write_status Enable or disable status graph creation and update. Enabled by default.
cache_size Maximum number of templates (used for file creation) kept on disk.
ignore_update_errors Ignore RRD files update errors (Broker 2.4 compatible behavior).

Example

<output>
  <type>rrd</type>
  <metrics_path>/var/lib/centreon/metrics</metrics_path>
  <status_path>/var/lib/centreon/status</status_path>
</output>

rrdcached

In large setups, the number of graphs handled by Centreon Broker can become very important and generate I/O massively. The rrdcached software was created to circumvent this issue as explained in its man page.

Due to current rrdcached limitations, Centreon Broker can only use it when local. This explains why you do not have any configuration tag for the rrdcached host.

SQL

Insert data in Centreon Broker database.

Type sql
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Configuration

Tag Description
db_type Type of the database (mysql, postgresql, oracle, ...).
db_host Database host.
db_port Database port.
db_user Database user.
db_password Password associated with db_user.
db_name Database name.
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits. This prevent database from not being updated due to lack of queries to fill the transaction.
check_replication Useful when using DB replication. Enable or disable replication check when connecting. Default is enabled.
cleanup_check_interval How often the cleanup thread should run. This thread cleans multiple tables of the database containing outdated data.
instance_timeout In seconds, how long the broker will wait for a poller before signaling the poller as unresponsive. 0 to disable this feature.
with_state_events Generate host/service state events. This is an experimental feature used to compute real-time BI information.

Example

<output>
  <type>sql</type>
  <db_type>mysql</db_type>
  <db_host>localhost</db_host>
  <db_port>3306</db_port>
  <db_user>centreon</db_user>
  <db_password>noertnec</db_password>
  <db_name>centreon_storage</db_name>
</output>

Storage

The storage module extract performance data from check results and insert them in a database. It also generate events for use by the modules rrd.

Type storage
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Configuration

Tag Description
interval Monitoring engine base interval (usually 60 seconds).
length RRD file length in seconds (ie. how much data your RRD file will contain). For 180 days (recommended), use 15552000 (180 * 24 * 60 * 60).
db_type Database type. One of db2, ibase, mysql, oracle, odbc, postgresql, sqlite, tds.
db_host Database host.
db_port Database port.
db_user Database user.
db_password Database password.
db_name Database name.
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits. This prevent database from not being updated due to lack of queries to fill the transaction.
check_replication Useful when using DB replication. Enable or disable replication check when connecting. Default is enabled.
rebuild_check_interval How often (in seconds) metrics should be checked for rebuild.
store_in_data_bin This can be used to avoid keeping performance data in the data_bin table. Warning: this will prevent you to rebuild RRD files.
insert_in_index_data Internal option used by Centreon to allow graphs to properly work on satellite (deported interface).

Example

<output>
  <type>storage</type>
  <interval>60</interval>
  <length>15552000</length>
  <db_type>oracle</db_type>
  <db_host>localhost</db_host>
  <db_port>1521</db_port>
  <db_user>centreon</db_user>
  <db_password>noertnec</db_password>
  <db_name>centreon_storage</db_name>
</output>

BAM

Monitoring endpoint

Compute Business Activity and Key Performance Indicator levels as well as meta-services. This is the improved version (as a Centreon Broker module) of the Centreon BAM extension.

Type bam
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Configuration of the monitoring endpoint

Tag Description
db_type Type of the database (mysql, postgresql, oracle, ...).
db_host Database host.
db_port Database port.
db_user Database user.
db_password Password associated with db_user.
db_name Centreon database name (usually centreon).
storage_db_name Storage database name (usually centreon_storage).
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits. This prevent database from not being updated due to lack of queries to fill the transaction.
check_replication Useful when using DB replication. Enable or disable replication check when connecting. Default is enabled.
command_file Centreon Engine external command file. This is used to provide check results on BAs and launch notifications as a consequence.

Reporting endpoint

Compute reporting information on BAs. This is only useful for use with Centreon BI to generate reports on Business Activities.

Type bam_bi
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Configuration of the reporting endpoint

Tag Description
db_type Type of the database (mysql, postgresql, oracle, ...).
db_host Database host.
db_port Database port.
db_user Database user.
db_password Password associated with db_user.
db_name Database name.
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits. This prevent database from not being updated due to lack of queries to fill the transaction.
check_replication Useful when using DB replication. Enable or disable replication check when connecting. Default is enabled.

Example

<output>
  <type>bam</type>
  <db_type>mysql</db_type>
  <db_host>localhost</db_host>
  <db_port>3306</db_port>
  <db_user>centreon</db_user>
  <db_password>noertnec</db_password>
  <db_name>centreon</db_name>
  <storage_db_name>centreon_storage</storage_db_name>
</output>

TCP

Probably one of the most used module. Provides network connectivity.

Type tcp
Layer(s) 1-3
Work on input Yes
Work on output Yes
Work on temporary No

Configuration

Tag Description
host Host to connect to. To have a server connection, do not use this tag.
one_peer_retention_mode Allow only one connection for a specific output. This option allow to keep retention on a server output (socket on a listen mode).
port Port on which Centreon Broker should listen (if no host is defined) or connect to.
protocol Choose your protocol (bbdo or ndo).
socket_write_timeout The number of microseconds to wait for data to be sent in the socket before the waiting timeout. If not present, or 0, the write will never time out.

Example

Input stream that waits for clients to connect on port 5668.

<input>
  <type>tcp</type>
  <port>5668</port>
  <protocol>bbdo</protocol>
</input>

Output stream that connects on host remotehost.tld on port 5668.

<output>
  <type>tcp</type>
  <host>remotehost.tld</host>
  <port>5668</port>
  <protocol>bbdo</protocol>
</output>

Output stream that connects on host remotehost.tld on port 5668 and allow single peer connection.

<output>
  <type>tcp</type>
  <host>remotehost.tld</host>
  <port>5668</port>
  <protocol>bbdo</protocol>
  <one_peer_retention_mode>1</one_peer_retention_mode>
</output>

BBDO

BBDO is a serialization layer introduced by Centreon Broker. Its name stands for Broker Binary Data Objects. This protocol was initially introduced to improve performance over the NDO protocol which was using data stringification. As its name suggests, BBDO is using raw binary data transfer which drastically reduce its CPU footprint on busy servers.

BBDO also supports feature negociation and can automatically encrypt and/or compress transmitted data. No configuration is required, whereas possible, to use compression and TLS modules.

Type N/A
Layer(s) 7
Work on input Yes
Work on output Yes
Work on temporary No

Configuration

Tag Description
protocol Must be set to bbdo.
negociation Enable or disable BBDO automatic feature negociation.

NDO

NDO is the historical but deprecated serialization layer for events. A serialization layer is required on some endpoints to reach the 7th layer (NDO is inserted at layer 7).

Type N/A
Layer(s) 7
Work on input Yes
Work on output Yes
Work on temporary No

Configuration

Tag Description
protocol Must be set to ndo.

Example

<input>
  <type>tcp</tcp>
  <port>5668</port>
  <protocol>ndo</protocol>
</input>

TLS

The TLS module is using GNU TLS to provide encryption. Encryption can work with two modes : either with provided certificates (and optionally authentication with a trusted CA) or by using anonymous mode. With this last mode, TLS just have to be enabled and encryption is configured by the software. It is this last mode that is used by the BBDO protocol.

Type N/A
Layer(s) 5
Work on input Yes
Work on output Yes

Options

Tag Description
tls Enable TLS protocol. It can either be used as anonymous (no public_cert nor private_key) or with appropriate settings (ca_certificate) used with certificate authentication.
private_key Private key.
public_cert Public certificate associated with private_key.
ca_certificate Trusted Certificate Authority certificate. If this parameter is set, the CA’s certificate is used to authenticate client connections which are denied if the peer key could not be validated.

Example

Output stream connecting to remotehost.tld using public.cert and private.key files.

<output>
  <type>tcp</type>
  <host>remotehost.tld</host>
  <port>5669</port>
  <public_cert>public.cert</public_cert>
  <private_key>private.key</private_key>
  <protocol>bbdo</protocol>
</output>

Input stream that authenticate clients using the trusted CA’s certificate (trusted_ca.cert).

<input>
  <type>tcp</type>
  <port>5669</port>
  <public_cert>public.cert</public_cert>
  <private_key>private.key</private_key>
  <ca_certificate>trusted_ca.cert</ca_certificate>
  <protocol>bbdo</protocol>
</input>

Compression

The compression module uses the zlib compression algorithm to reduce the size of data transmitted by Centreon Broker. Typical compression ratio range from 2:1 to 5:1.

Type N/A
Layer(s) 6
Work on input Yes
Work on output Yes

Options

The compression can be configured with the tags defined in the table below.

Tag Description
compression Set it to 1 to enable compression. This is the sole mandatory parameter.
compression_level Level of compression from 0 (no compression) to 9 (best compression). Defaults to -1 which is zlib’s default compression level.
compression_buffer Size in bytes of the compression buffer. The biggest the buffer is, the best is the compression. However the latency increase along with the buffer size.

Example

<output>
  <type>tcp</type>
  <host>localhost</host>
  <port>5668</port>
  <compression>1</compression>
  <compression_level>4</compression_level>
  <compression_buffer>5000</compression_buffer>
</output>

Statistics

The statistics (stats) module was created to provide information about the ongoing processing of Centreon Broker. You find various statistics such as the number of events processed by second per input or output, the number of queued events, connected peers, loaded modules, ...

This is a global module that do not apply to a specific endpoint.

Type N/A
Layer(s) N/A
Work on input No
Work on output No
Work on temporary No

Configuration

A stats node must be defined right under the root node of the XML configuration file. This node can then contain the following tags.

Tag Description
fifo The FIFO file from which you can read the statistics.
remote Send statistics information from monitoring engine or broker over the networks (This feature need centreon-broker >= 2.7).

Example

<stats>
  <fifo>/var/lib/centreon-broker/central-module.stats</fifo>
</stats>

You can then read the file with a simple cat command.

$> cat /var/lib/centreon-broker/central-module.stats
module /usr/share/centreon/lib/centreon-broker/50-tcp.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/20-correlation.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/10-neb.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/60-compression.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/20-storage.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/80-sql.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/70-rrd.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/50-local.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/80-ndo.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/05-stats.so
state=loaded

module /usr/share/centreon/lib/centreon-broker/50-file.so
state=loaded

output poller-module
state=connected
queued_events=0
last event at=1358863864
event processing speed=160.3 events/s
last connection attempt=1358862546
last connection success=1358862546

Since Centreon-Broker 2.7 we can send statistics over the network with the node “remote”. This node can then contain the following tags.

Tag Description
dumper_tag Tag name of the dumper output to write statistics.
interval Interval in seconds to dump statitics.
metrics This node contain informations about statistics to write (see below). All stats into metrics create RRD files.

This table describe the node metrics.

Tag Description
host Host id of the metric.
service Service contain service id and servce name of the metric.

This table describe the node service.

Tag Description
id Service id of the metric.
name Service name.

This is list of available services for NEB module.

Name Description Perfdata
active_host_execution_time Active host check execution time (ms) avg, min, max
active_host_latency Active host check latency (ms) avg, min, max
active_hosts_last Number of active host checks in last minutes active_hosts_last_1, active_hosts_last_5, active_hosts_last_15, active_hosts_last_60
active_host_state_change Active host check % state change avg, min, max
active_service_execution_time Active service check execution time (ms) avg, min, max
active_service_latency Active service check latency (ms) avg, min, max
active_services_last Number of active service checks in last minutes active_services_last_1, active_services_last_5, active_services_last_15, active_services_last_60
active_service_state_change Active service check % state change avg, min, max
command_buffers External command buffer informations used, high, total
hosts_actively_checked Total of hosts actively checked hosts_actively_checked
hosts_checked Total of hosts checked hosts_checked
hosts_flapping Number of host in flapping hosts_flapping
hosts Total hosts state informations up, down, unreachable
hosts_in_downtime Number of host in downtime hosts_in_downtime
hosts_passively_checked Total of hosts passively checked hosts_passively_checked
hosts_scheduled Total of scheduled hosts hosts_scheduled
passive_host_latency Passive host check latency (ms) avg, min, max
passive_hosts_last Number of passive host checks in last minutes passive_hosts_last_1, passive_hosts_last_5, passive_hosts_last_15, passive_hosts_last_60
passive_host_state_change Passive host check % state change avg, min, max
passive_service_latency Passive service check latency (ms) avg, min, max
passive_services_last Number of passive service checks in last minutes passive_services_last_1, passive_services_last_5, passive_services_last_15, passive_services_last_60
passive_service_state_change Passive service check % state change avg, min, max
services_actively_checked Total of services actively checked services_actively_checked
services_checked Total of services checked services_checked
services_flapping Number of service in flapping services_flapping
services Total services state informations ok, warning, critical, unknown
services_in_downtime Number of service in downtime services_in_downtime
services_passively_checked Total of services passively checked services_passively_checked
services_scheduled Total of scheduled services services_scheduled
total_hosts Total number of hosts total_hosts
total_host_state_change Total host check % state change avg, min, max
total_services Total number of services total_services
total_service_state_change Total service check % state change avg, min, max

Example

<stats>
  <remote>
    <dumper_tag>CentralBroker</dumper_tag>
    <interval>300</interval>
    <metrics>
      <host>1</host>
      <service>
        <id>1</id>
        <name>active_host_execution_time</name>
      </services>
      <service>
        <id>2</id>
        <name>active_host_latency</name>
      </services>
    </metrics>
  </remote>
</stats>

Correlation

The correlation module provide basic correlation features to Centreon Broker.

This is a global module that do not apply to a specific endpoint.

Type N/A
Layer(s) N/A
Work on input No
Work on output No
Work on temporary No

Configuration

A correlation node must be defined right under the root node of the XML configuration file. This node can then contain the tags described in the table below.

Tag Description
file The XML configuration file containing host and service definitions along with parenting and dependencies definitions.
retention The XML retention file. This file is written by Centreon Broker so that the correlation engine does not forget the current states of the hosts and services across Broker restarts.
passive Enable passive mode. In this mode, the correlator update the internal state with correlation events. No events are emits in passive mode.

Example

<correlation>
  <file>/etc/centreon-broker/correlation.cfg</file>
  <retention>/var/lib/centreon-broker/correlation.sav</retention>
</correlation>

Dumper

The dumper module is used to transfer files between Centreon Broker instances.

Dumper endpoint

The dumper endpoint is used to write files on the disk. It can filter data by tagname. A dumper write only data with the same tagname as his.

Type dumper
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Configuration of the dumper endpoint

Tag Description
tagname The tag name to write file. Tag name allow filtering.
path The path of the file to write data. You can use some macros in the path. Available macros are $INSTANCEID$ (the instance ID of the Centreon Broker instance that generated the dump event) and $FILENAME$ (the original file name of the file that generated the dump event).

FIFO dump endpoint

This endpoint transfer the content of a FIFO file. Each time a line is written to the FIFO, it is transfered.

Type dump_fifo
Layer(s) 1-7
Work on input Yes
Work on output No
Work on temporary No

Configuration of the FIFO dump endpoint

Tag Description
tagname The tag name to write file. This is used by dumper endpoints to perform filtering.
path The path of the FIFO file.

Directory dump endpoint

This endpoint transfer the content of a directory. Each time a file is modified it is transfered.

Type dump_dir
Layer(s) 1-7
Work on input Yes
Work on output No
Work on temporary No

Configuration of the directory dump endpoint

Tag Description
tagname The tag name to write file. This is used by dumper endpoints to perform filtering.
path The path of the directory to watch for changes.

Example

<input>
  <type>dump_fifo</type>
  <tagname>engine-extcmd-1</tagname>
  <path>/var/lib/centreon/poller-1.cmd</path>
</input>

<input>
  <type>dump_dir</type>
  <tagname>engine-cfg-1</tagname>
  <path>/var/lib/centreon/cfg/1</path>
</input>

<output>
  <type>dumper</type>
  <tagname>CentralBroker</tagname>
  <path>/var/lib/centreon-broker/$INSTANCEID$.stats</path>
</output>

<output>
  <type>dumper</type>
  <tagname>engine-extcmd-1</tagname>
  <path>/var/lib/centreon-engine/rw/centengine.cmd</path>
</output>

<output>
  <type>dumper</type>
  <tagname>engine-cfg-1</tagname>
  <path>/etc/centreon-engine/$FILENAME$</path>
</output>

InfluxDB

Warning

This module is experimental.

This module fills an InfluxDB instance with metrics. It uses the storage engine as its performance data source.

Type influxdb
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary No

Description

Users of this module can make their own storage scheme to suite their needs. The InfluxDB module will provide information on a particular metric using a macro system. Each macro must be written between two dollar-sign (like $MACRO$). If one wishes to write a dollar-sign he must write two continguous dollar-signs ($$).

Configuration

Tag Description
db_host Database host.
db_port Database port. Default 8086.
db_user Database user.
db_password Password associated with db_user.
db_name Database name.
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits in seconds. This prevent database from not being updated due to lack of queries to fill the transaction. Default to 1s.
status_timeseries Name of the time series for statuses. Macros accepted.
status_column Define one status column. Each InfluxDB endpoint can have multiple status columns. Macros accepted.
status_column -> name Name of the column. Macros accepted.
status_column -> value Value of the column. Macros accepted.
status_column -> is_tag Boolean value that is set to true if the current column is a tag. Defaults to false.
status_column -> type Type of the column. Can be either ‘string’ or ‘number’. Default to ‘string’.
metrics_timeseries Name of the time series for metrics. Macros accepted.
metrics_column Define one metrics column. Each InfluxDB endpoint can have multiplex metrics columns. Macros accepted.
metrics_column -> name Name of the column. Macros accepted.
metrics_column -> value Value of the column Macros accepted.
metrics_column -> is_tag Boolean value that is set to true if the current column is a tag. Defaults to false.
metrics_column -> type Type of the column. Can be either ‘string’ or ‘number’. Default to ‘string’.

Available macros are described in the following table.

Macro Description For statuses For metrics
INSTANCE Instance (poller) name. X X
INSTANCEID Instance (poller) ID. X X
HOST Host. X X
HOSTID Host ID. X X
SERVICE Service. X X
SERVICEID Service ID. X X
INDEXID Index of the host / service within Centreon’s index_data table. X X
METRIC Metric name.   X
METRICID ID of the metric within Centreon’s metrics table.   X

Example

<output>
  <type>influxdb</type>
  <db_host>localhost</db_host>
  <db_port>8086</db_port>
  <db_user>centreon</db_user>
  <db_password>noertnec</db_password>
  <db_name>metrics</db_name>

  <!-- All statuses will be stored in the 'status' time series. -->
  <status_timeseries>status</status_timeseries>
  <status_column>
    <name>time</name>
    <value>$TIME$</value>
  </status_column>
  <status_column>
    <name>value</name>
    <value>$VALUE$</value>
  </status_column>
  <status_column>
    <name>index_id</name>
    <value>$INDEXID$</value>
    <is_tag>1</is_tag>
  </status_column>

  <!-- Each separate metric will be stored in a time series that has
       the same name as the metric name. -->
  <metrics_timeseries>$METRIC$</metrics_timeseries>
  <metrics_column>
    <name>time</name>
    <value>$TIME$</value>
  </metrics_column>
  <metrics_column>
    <name>value</name>
    <value>$VALUE$</value>
  </metrics_column>
  <metrics_column>
    <name>metric_id</name>
    <value>$METRICID$</value>
    <is_tag>1</is_tag>
  </metrics_column>
</output>

Graphite

Warning

This module is experimental.

This module fills a Graphite instance with metrics. It uses the storage engine as its performance data source.

Type graphite
Layer(s) 1-7
Work on input No
Work on output Yes
Work on temporary NO

Description

The module uses only the plaintext protocol.

Configuration

Tag Description
metric_naming Naming hierarchy within Graphite. This defaults to centreon.metrics.$METRICID$ where $METRICID$ is the metric’s ID within Centreon’s metrics table. Complete available variables are $INSTANCE$ (name of the poller), $INSTANCEID$, $HOST$ (the host name), $HOSTID$, $SERVICE$ (the service description), $SERVICEID$, $METRIC$ (the metric name), $METRICID$ and $INDEXID$.
status_naming Naming hierarchy within Graphite. This defaults to centreon.statuses.$INDEXID$ where $INDEXID$ is the index ID of the host / service within Centreon’s index_data table. Complete available variables are $INSTANCE$, $INSTANCEID$, $HOST$, $HOSTID$, $SERVICE$, $SERVICEID$ and $INDEXID$.
db_host Database host.
db_port Database port. Default to 80.
db_user Database user. Default is empty (no authentication).
db_password Password associated with db_user. Default is empty (no authentication).
queries_per_transaction Number of queries per transaction. Set to 1 or below to disable transactions. Default to 1.
read_timeout When using transactions, maximum time between commits in seconds. This prevent database from not being updated due to lack of queries to fill the transaction. Default to 1s.