This source reads and automatically parses the Linux audit logs. You can override the file name using the filename() parameter and the prefix for the created name-value pairs using the prefix() parameter. Any additional parameters are passed to the underlying file source.
NOTE: Most recent Linux distributions enable Security-Enhanced Linux (SELinux) or AppArmor as a security measure. If enabled, these technologies might disable access to the Linux Audit log file by default. Consult their manuals to enable Linux Audit log access for syslog-ng PE.
Declaration
linux-audit(options);
Example: Using the linux-audit() driver
source s_auditd {
linux-audit(
prefix("test.")
hook-commands(
startup("auditctl -w /etc/ -p wa")
shutdown("auditctl -W /etc/ -p wa")
)
);
};
The linux-audit() driver has the following options:
filename()
Description: The log file of linux-audit. The syslog-ng PE application reads the Linux audit logs from this file.
prefix()
| Synopsis: |
prefix() |
| Default: |
.auditd. |
Description: Insert a prefix before the name part of the parsed name-value pairs to help further processing. For example:
-
To insert the my-parsed-data. prefix, use the prefix(my-parsed-data.) option.
-
To refer to a particular data that has a prefix, use the prefix in the name of the macro, for example, ${my-parsed-data.name}.
-
If you forward the parsed messages using the IETF-syslog protocol, you can insert all the parsed data into the SDATA part of the message using the prefix(.SDATA.my-parsed-data.) option.
Names starting with a dot (for example, .example) are reserved for use by syslog-ng PE. Note that if you use an empty prefix (prefix("")) or one starting with a dot, syslog-ng PE might replace the original value of an existing macro (note that only soft macros can be overwritten, see Hard versus soft macros for details). To avoid such problems, use a prefix when naming the parsed values, for example, prefix(my-parsed-data.)
From version 7.0.21, syslog-ng Premium Edition (syslog-ng PE) can collect messages from a Microsoft SQL database (using an mssql() source), or from an Oracle database (using an oracle() source). For backward compatibility with syslog-ng PE 6LTS, sql() source is also available.
NOTE: For new deployments, One Identity recommends using a dedicated mssql() or oracle() source instead of a generic sql() source.
Currently, the Microsoft SQL (MSSQL) and Oracle databases are supported.
For more information, see Upgrading the sql() source of syslog-ng PE.
Limitations
-
The Oracle database is not supported on Oracle Linux 6 and RedHat 6.
-
The sql() source driver does not monitor rotated tables. As a result, every source can follow only one table.
-
Timestamps with timezone are not supported. The syslog-ng PE application will retrieve the timestamps from these columns, but without the timezone information.
-
The sql() source driver ignores the log-msg-size() option (that is, messages read from the SQL database can be longer than the maximal message length set in the log-msg-size() option).
-
There is an ID column that is the monotonically increasing unique ID of the monitored table. It is not possible to use more than one ID columns as a complex ID.
Declaration
The following examples can be used as a template for declaring the mssql(), oracle(), or sql() sources in your configuration.
Example: declaration for the mssql() source
You can use the following example as a template for declaring the mssql() source in your configuration:
source s_mssql {
mssql(
host(host)
username(username)
password(pass)
port(1433)
database(test_db)
table(table_name)
uid_column(uid)
datetime_column(datetime)
message_template('uid: ${.sql.uid} datetime: ${.sql.datetime}')
);
};
Example: declaration for the oracle() source
You can use the following example as a template for declaring the oracle() source in your configuration:
source s_oracle {
oracle(
host(host)
username(username)
password(pass)
port(1433)
database(test_db)
table(table_name)
uid_column(uid)
datetime_column(datetime)
message_template('uid: ${.sql.uid} datetime: ${.sql.datetime}')
);
};
Example: declaration for the sql() source
You can use the following example as a template for declaring the sql() source in your configuration:
source s_sql {
sql(
type(mssql)
host(host)
username(username)
password(pass)
port(1433)
database(test_db)
table(table_name)
uid_column(uid)
datetime_column(datetime)
message_template('uid: ${.sql.uid} datetime: ${.sql.datetime}')
);
};
Required parameters
The mssql(), oracle(), and sql() drivers have the following required parameters in common:
The oracle() driver has an additional required parameter: start-uid().
The sql() driver has an additional required parameter: type().
The mssql(), oracle(), and sql() drivers have the following options.
columns()
| Type: |
string list |
| Default: |
empty |
Description: The list of the name of the columns that will be queried. The default value is empty, meaning that all of the columns will be queried.
Example
columns("id","date","message")
connection-timeout()
| Type: |
nonnegative integer |
| Default: |
5 |
Description: Optional. The syslog-ng PE application waits connection-timeout() seconds for any request to complete. If a request does not succeed in the given time, syslog-ng PE will disconnect, and tries to reconnect after time-reopen seconds. 0 value means infinite timeout.
connect-query()
Description: The SQL-like statement which is executed after syslog-ng PE has successfully connected to the database.
|
|
Caution:
Hazard of data loss!
The syslog-ng PE application does not validate or limit the contents of customized queries. Consequently, queries performed with a user with write-access can potentially modify or even harm the database. Use customized queries with care, and only on your own responsibility. |
|
|
Caution:
Using unique column names in mssql() fetch-query() results is mandatory from syslog-ng PE 7.0.33. Explicit usage of "AS" SQL keyword could help to achieve this. Ignoring it may raise an error in mssql() source driver and queries will be dropped. |
Example: sample connect-query() statement
The following example is a sample connect-query() statement:
connect-query("SET COLLATION_CONNECTION='utf8_general_ci'")
database()
| Type: |
string |
| Default: |
logs |
Description: Name of the database that stores the logs. Macros cannot be used in database name.
date-column(col_name, [format])
| Type: |
date, string |
| Default: |
|
Description: The column containing the date of the logrecord. The format value has to be in strptime format. For details, see strptime(3) - Linux man page.
datetime-column(col_name, [format])
Description: The column containing the timestamp. If the type is int, it is considered to contain a UNIX timestamp. The format value is required if the type is string, and has to be in strptime format. For details, see strptime(3) - Linux man page.
The following column types are supported:
-
oracle(): timestamp, int
-
mssql(): datetime, int
Example: sample datetime-column(col_name, [format])
The following example is a sample datetime-column(col_name, [format]) column containing the timestamp:
columns("id","date","message")datetime("timestampcol", "%Y-%m-%d")
default-facility()
| Type: |
facility string |
| Default: |
user |
Description: This parameter assigns a facility value to the messages received from the sql() source.
default-severity()
| Type: |
severity string |
| Default: |
notice |
Description: This parameter assigns a severity level to the messages received from the sql() source.
fast-follow-mode()
| Type: |
yes | no |
| Default: |
yes |
Description: If set to yes, syslog-ng PE reads the database table as fast as possible, until it reaches the last record. After this, it will execute only one query in follow-freq() time. If it is set to no, syslog-ng PE executes only one query in follow-freq() time.
fetch-query()
| Type: |
string |
| Default: |
Default value is generated in running time of syslog-ng PE |
Description: The SQL-like statement used to collect the records from the database.
NOTE: If this parameter is defined, syslog-ng PE does not check or validate it whether it is correct. Ensure that the customized statements are correct.
For details on customizing queries, see Customizing mssql() queries.
|
|
Caution:
Hazard of data loss!
The syslog-ng PE application does not validate or limit the contents of customized queries. Consequently, queries performed with a user with write-access can potentially modify or even harm the database. Use customized queries with care, and only on your own responsibility. |
|
|
Caution:
Using unique column names in mssql() fetch-query() results is mandatory from syslog-ng PE 7.0.33. Explicit usage of "AS" SQL keyword could help to achieve this. Ignoring it may raise an error in mssql() source driver and queries will be dropped. |
Example: sample fetch-query() statement
The following example is a sample fetch-query() statement:
fetch-query("SELECT * FROM $table WHERE id > $last_read_uid AND test_logs.log LIKE '%ERROR%' ORDER BY $uid")
The default fetch-query() statements for the mssql() and oracle() sources are the following:
-
mssql():
SELECT TOP $fetch_limit $columns FROM $table WHERE $uid > '$last_read_uid' ORDER BY $uid
-
oracle():
SELECT $columns FROM (SELECT * FROM $table WHERE $uid > $last_read_uid ORDER BY $uid) WHERE rownum <= $fetch_limit
follow-freq()
| Type: |
number(seconds) |
| Default: |
If time-reopen() is set to a value other than the default 60, the value of time-reopen(). Otherwise 60. |
Description: The syslog-ng PE application checks whether the SQL source changed every time the follow-freq() interval (in seconds) has elapsed. Floating-point numbers (for example, 1.5) can be used as well.
host()
| Type: |
hostname or IP address |
| Default: |
n/a |
Description: Hostname of the database server.
host-template()
| Type: |
string |
| Default: |
empty string |
Description: The template for defining the HOST part of the message. If the host-template() option is not specified, the value of the host() option will be used in the HOST part of the message.
log-fetch-limit()
Description: The maximum number of messages fetched from a source during a single poll loop.
log-iw-size()
| Type: |
number (messages) |
| Default: |
100 |
Description: The size of the initial window, this value is used during flow control.
login-timeout()
| Type: |
nonnegative integer |
| Default: |
5 |
Description: Optional. The syslog-ng PE application waits login-timeout() seconds for the login request to complete. If the request does not succeed in the given time, syslog-ng PE will try to reconnect after time-reopen() seconds. 0 value means infinite timeout.
max-uid-query()
| Type: |
string |
| Default: |
SELECT max($uid) FROM $table |
Description: Used for retrieving the ID of the last row (that is, the last row of the last fetch) from the database source.
message-template()
Description: The alias of the template() parameter.
password()
| Type: |
string |
| Default: |
n/a |
Description: Password of the database user.
port()
| Type: |
number (port number) |
| Default: |
1433 TCP for MSSQL, 1521 for Oracle |
Description: The port number to connect to.
prefix()
| Type: |
string |
| Default: |
.sql. |
Description: This prefix will be added to the name of the macros created from the database columns.
Example: sample prefix() - database column name - macro name correlation
The prefix() and database column name affect the name of the macro created from the database column.
For example, if a database column is called column1, and the prefix option is set as prefix("customprefix."), the macro for the column will be called customprefix.column1.
program-template()
| Type: |
string |
| Default: |
empty string |
Description: The template for defining the PROGRAM part of the message. If not specified, the PROGRAM message part will be empty.
start-uid()
Description:
- mssql(): Optional.
- oracle(): Mandatory. If start-uid() is specified, syslog-ng PE will only query entries with a strictly larger uid. After the first successfully fetched message, syslog-ng PE does not use this option anymore, even after restart or reload. If start-uid() is not specified, syslog-ng PE will read only the new records.
table()
Description: The name of the monitored table. Only a single literal name is accepted, macros cannot be used in the name of the table. Monitoring rotated tables is not supported.
table-init-query()
Description: The SQL-like statement which is executed before fetching the first batch of records.
For details on customizing MSSQL queries, see Customizing mssql() queries.
|
|
Caution:
Hazard of data loss!
The syslog-ng PE application does not validate or limit the contents of customized queries. Consequently, queries performed with a user with write-access can potentially modify or even harm the database. Use customized queries with care, and only on your own responsibility. |
|
|
Caution:
Using unique column names in mssql() fetch-query() results is mandatory from syslog-ng PE 7.0.33. Explicit usage of "AS" SQL keyword could help to achieve this. Ignoring it may raise an error in mssql() source driver and queries will be dropped. |
tags()
Description: Label the messages received from the source with custom tags. Tags must be enclosed between double quotes. When adding multiple tags, separate them with commas, for example, tags("dmz", "router").
template()
Description: The template of the message (${MSG}) to be generated. If not specified, the following template will be used: "$(format-welf --key <prefix>*)") where the <prefix> is the value of the prefix() option. This template converts the retrieved records into the WebTrends Enhanced Log file Format (WELF).
NOTE: The format-welf function does not keep the order of columns between queries.
Example: sample queries' results for a table with two columns using the template() option
The following examples show the results of the first, and second queries when using the default template for a table that has two columns (id and message):
The result of the first query is the following:
'.sql.id=12 .sql.message="test message"'
The result of the second query can be:
'.sql.message="test message" .sql.id=12 '
time-column(col_name, [format])
| Type: |
time, string |
| Default: |
|
Description: The column containing the time of the logrecord. The format value has to be in strptime format.
time-reopen()
| Type: |
number (seconds) |
| Default: |
60 |
Description: The time to wait in seconds before a broken connection is reestablished.
tls()
| Type: |
tls options |
| Default: |
none |
Description: The tls() configuration block enables TLS encryption for MSSQL Server.
NOTE: The tls() configuration block used for the network() source is similar in functionality, but there is no 100% feature parity between them. Consider the details for each tls() configuration block option before using them with your mssql() source.
The tls() configuration block has the following options:
-
ca-file()
Type: Filename
Default: none
Description: The name of a file that contains an X.509 CA certificate (or a certificate chain) in PEM format. The syslog-ng PE application uses this certificate to validate the certificate of the MSSQL server.
-
peer-verify()
Type: yes | no
Default: yes
Description: In syslog-ng PE version 7.0.28, this option only enables an extra check for the server CN matching against the hostname (configured in syslog-ng PE via the host() option).
NOTE: This option is only effective if you also specify a ca-file() in your configuration block.
-
ssl-options()
Type: comma-separated list
Possible values: no-tlsv1 | none
Default: none
Description: This option allows disabling TLS protocols. Consider that it only disables the current version, but not older versions.
NOTE: The default none value does not restrict TLS protocols, but uses the default TLS value set for the mssql() source ( 1.0, 1.2).
type()
| Type: |
mssql, or oracle |
| Default: |
|
Description: Specifies the type of the database. Only use it with the generic sql() source.
uid-column()
| Type: |
string or string type(string)
Supported column types are: integer, character type, datetime, datetime2 |
| Default: |
|
Description: The monotonically increasing unique ID of the monitored table (for example, auto_increment). This column must be a type where the greater (>) operation is interpreted.
The second optional parameter is a type hint (for example, uid-column(uid' type('DATE')). If provided, syslog-ng PE uses the given string in the SQL queries.
username()
| Type: |
string |
| Default: |
n/a |
Description: Name of the database user.
use-syslogng-pid()
| Type: |
yes | no |
| Default: |
no |
Description: If the value of this option is yes, then the pid value of the message will be overridden with the pid of the running syslog-ng PE process.
use-tnsnames()
| Type: |
yes | no |
| Default: |
For sql(): yes
For oracle(): no |
Description: Optional. This option is used only in case of oracle(), or sql() source with type("oracle") option. If set to yes, syslog-ng PE will use the connection parameters defined in the tnsnames.ora configuration file.
