1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
|
Lookup Tables
=============
Lookup tables are a powerful construct to obtain *class* information based
on message content. It works on top of a data-file which maps key (to be looked
up) to value (the result of lookup).
The idea is to use a message properties (or derivatives of it) as an index
into a table which then returns another value. For example, $fromhost-ip
could be used as an index, with the table value representing the type of
server or the department or remote office it is located in.
This can be emulated using if and else-if stack, but implementing it as a
dedicated component allows ``lookup`` to be made fast.
The lookup tables itself exists in a separate data file (one per
table). This file is loaded on Rsyslog startup and when a reload is requested.
There are different types of lookup tables (identified by "type" field in json data-file).
Types
^^^^^
string
------
The key to be looked up is an arbitrary string.
**Match criterion**: The key must be exactly equal to index from one of the entries.
array
-----
The value to be looked up is an integer number from a consecutive set.
The set does not need to start at zero or one, but there must be no number missing.
So, for example ``5,6,7,8,9`` would be a valid set of index values, while ``1,2,4,5`` would
not be (due to missing ``3``).
**Match criterion**: Looked-up number(key) must be exactly equal to index from one of the entries.
sparseArray
-----------
The value to be looked up is an integer value, but there may be gaps inside the
set of values (usually there are large gaps). A typical use case would be the
matching of IPv4 address information.
**Match criterion**: A match happens on the first index that is less than or equal to the looked-up key.
Note that index integer numbers are represented by unsigned 32 bits.
Lookup Table File Format
^^^^^^^^^^^^^^^^^^^^^^^^
Lookup table files contain a single JSON object. This object consists of a header and a table part.
**Header**
The header is the top-level json.
It has parameters "version", "nomatch", and "type".
Parameters:
**version** <number, default: 1> : Version of table-definition format (so improvements in the future can be managed in a backward compatible way).
**nomatch** <string literal, default: ""> : Value to be returned for a lookup when match fails.
**type** <*string*, *array* or *sparseArray*, default: *string*> : Type of lookup-table (controls how matches are performed).
**Table**
This must be an array of elements, even if only a single value exists (for obvious reasons,
we do not expect this to occur often). Each array element must contain two fields "index"
and "value".
This is a sample of how an ip-to-office mapping may look like:
::
{ "version" : 1,
"nomatch" : "unk",
"type" : "string",
"table" : [
{"index" : "10.0.1.1", "value" : "A" },
{"index" : "10.0.1.2", "value" : "A" },
{"index" : "10.0.1.3", "value" : "A" },
{"index" : "10.0.2.1", "value" : "B" },
{"index" : "10.0.2.2", "value" : "B" },
{"index" : "10.0.2.3", "value" : "B" }]}
Note: In the example above, if a different IP comes in, the value "unk" is returned thanks to the nomatch parameter in the first line.
Lookup tables can be accessed via the ``lookup()`` built-in function. A common usage pattern is to set a local variable to the lookup result and later use that variable in templates.
Lookup-table configuration
^^^^^^^^^^^^^^^^^^^^^^^^^^
Lookup-table configuration involves a **two part setup** (definition and usage(lookup)), with an optional third part,
which allows reloading table using internal trigger.
lookup_table(name="<table>" file="</path/to/file>"...) (object)
---------------------------------------------------------------
**Defines** the table(identified by the table-name) and allows user to set some properties that control behavior of the table.
::
lookup_table(name="msg_per_host")
Parameters:
**name** <string literal, mandatory> : Name of the table.
**file** <string literal, file path, mandatory> : Path to external json database file.
**reloadOnHUP** <on|off, default: on> : Whether or not table should be reloaded when process receives HUP signal.
A definition setting all the parameters looks like:
::
lookup_table(name="host_bu" file="/var/lib/host_billing_unit_mapping.json" reloadOnHUP="on")
lookup("<table>", <expr>) (function)
------------------------------------
**Looks up** and returns the value that is associated with the given key (passed as <variable>)
in lookup table identified by table-name. If no match is found (according to table-type
matching-criteria specified above), the "nomatch" string is returned (or an empty string if it is not defined).
Parameters:
**name** <string literal, mandatory> : Name of the table.
**expr** <expression resulting in string or number according to lookup-table type, mandatory> : Key to be looked up.
A ``lookup`` call looks like:
::
set $.bu = lookup("host_bu", $hostname);
if ($.bu == "unknown") then {
....
}
Some examples of different match/no-match scenarios:
**string table**:
::
{ "nomatch" : "none",
"type" : "string",
"table":[
{"index" : "foo", "value" : "bar" },
{"index" : "baz", "value" : "quux" }]}
Match/no-Match behaviour:
====== ==============
key return
====== ==============
foo bar
baz quux
corge none
====== ==============
**array table**:
::
{ "nomatch" : "nothing",
"type" : "array",
"table":[
{"index" : 9, "value" : "foo" },
{"index" : 10, "value" : "bar" },
{"index" : 11, "value" : "baz" }]}
Match/no-Match behaviour:
====== ==============
key return
====== ==============
9 foo
11 baz
15 nothing
0 nothing
====== ==============
**sparseArray table**:
::
{ "nomatch" : "no_num",
"type" : "sparseArray",
"table":[
{"index" : "9", "value" : "foo" },
{"index" : "11", "value" : "baz" }]}
Match/no-Match behaviour:
====== ==============
key return
====== ==============
8 no_num
9 foo
10 foo
11 baz
12 baz
100 baz
====== ==============
reload_lookup_table("<table>", "<stub value>") (statement)
----------------------------------------------------------
**Reloads** lookup table identified by given table name **asynchronously** (by internal trigger, as opposed to HUP).
This statement isn't always useful. It needs to be used only when lookup-table-reload needs to be triggered in response to
a message.
Messages will continue to be processed while table is asynchronously reloaded.
Note: For performance reasons, message that triggers reload should be accepted only from a trusted source.
Parameters:
**name** <string literal, mandatory> : Name of the table.
**stub value** <string literal, optional> : Value to stub the table in-case reload-attempt fails.
A ``reload_lookup_table`` invocation looks like:
::
if ($.do_reload == "y") then {
reload_lookup_table("host_bu", "unknown")
}
Implementation Details
^^^^^^^^^^^^^^^^^^^^^^
The lookup table functionality is implemented via efficient algorithms.
The string and sparseArray lookup have O(log(n)) time complexity, while array lookup is O(1).
To preserve space and, more important, increase cache hit performance, equal data values are only stored once,
no matter how often a lookup index points to them.
|