exec - Fluentd

Search…

exec

The out_exec TimeSliced Output plugin passes events to an external program. The program receives the path to a file containing the incoming events as its last argument. The file format is tab-separated values (TSV) by default.
It is included in Fluentd's core.
Example Configuration
1
<match pattern>
2
  @type exec
3
  command cmd arg arg
4
  <format>
5
    @type tsv
6
    keys k1,k2,k3
7
  </format>
8
  <inject>
9
    tag_key k1
10
    time_key k2
11
    time_format %Y-%m-%d %H:%M:%S
12
  </inject>
13
</match>
Copied!
Please see the Configuration File article for the basic structure and syntax of the configuration file.
Example: Running FizzBuzz Against Data Stream
This example illustrates how to run FizzBuzz with out_exec.
We assume that the input file is specified by the last argument in the command line (ARGV[-1]). The following script fizzbuzz.py runs FizzBuzz against the new-line delimited sequence of natural numbers (1, 2, 3...) and writes the output to foobar.out:
1
#!/usr/bin/env python
2
import sys
3
input = file(sys.argv[-1])
4
output = file("foobar.out", "a")
5
for line in input:
6
    fizzbuzz = int(line.split("\t")[0])
7
    s = ''
8
    if fizzbuzz%3 == 0:
9
        s += 'fizz'
10
    if fizzbuzz%5 == 0:
11
        s += 'buzz'
12
    if len(s) > 0:
13
        output.write(s+"\n")
14
    else:
15
        output.write(str(fizzbuzz)+"\n")
16
output.close
Copied!
Note that this program is written in Python. For out_exec (as well as out_exec_filter and in_exec), the program can be written in any language, not just Ruby.
Then, configure Fluentd as follows:
1
<source>
2
  @type forward
3
</source>
4
<match fizzbuzz>
5
  @type exec
6
  command python /path/to/fizzbuzz.py
7
  <buffer>
8
    @type file
9
    path /path/to/buffer_path
10
    flush_interval 5s # for debugging/checking
11
  </buffer>
12
  <format>
13
    @type tsv
14
    keys fizzbuzz
15
  </format>
16
</match>
Copied!
The @type tsv and keys fizzbuzz in <format> tells Fluentd to extract the fizzbuzz field and output it as TSV. This simple example has a single key, but you can of course extract multiple fields and use format json to output newline-delimited JSON.
The intermediate TSV is at /path/to/buffer_path, and the command python /path/to/fizzbuzz.py /path/to/buffer_path is run. This is why in fizzbuzz.py, it is reading the file at sys.argv[-1].
If you start Fluentd and run this command:
1
$ for i in `seq 15`; do echo "{\"fizzbuzz\":$i}" | fluent-cat fizzbuzz; done
Copied!
Then, after 5 seconds, you get a file named foobar.out:
1
$ cat foobar.out
2
1
3
2
4
fizz
5
4
6
buzz
7
fizz
8
7
9
8
10
fizz
11
buzz
12
11
13
fizz
14
13
15
14
16
fizzbuzz
Copied!
Supported Modes
Asynchronous
See Output Plugin Overview for more details.
Plugin Helpers
​inject​
​formatter​
​compat_parameters​
​child_process​
Parameters
​Common Parameters​
@type
The value must be exec.
command
type
default
version
string
Nothing
0.14.0
The command (program) to execute. The exec plugin passes the path of flushed buffer chunk as the last argument.
If you set command parameter like this:
1
command cmd arg arg
Copied!
The actual command execution is:
1
cmd arg arg /path/to/file
Copied!
If cmd does not exist in PATH, you need to specify the absolute path, e.g. /path/to/cmd.
command_timeout
type
default
version
time
270
0.14.9
Command (program) execution timeout.
<format> Section
See Format Section for more details.
@type
type
default
version
string
tsv
0.14.9
The format used to map the incoming events to the program input.
Overwrites the default value in this plugin.
<inject> Section
See Inject Section for more details.
time_type
type
default
version
string
string
0.14.9
Overwrites the default value in this plugin.
localtime
type
default
version
bool
false
0.14.9
Overwrites the default value in this plugin.
<buffer> Section
See Buffer Section for more details.
delayed_commit_timeout
type
default
version
time
300
0.14.9
Overwrites the default value in this plugin.
If this article is incorrect or outdated, or omits critical information, please let us know. Fluentd is an open-source project under Cloud Native Computing Foundation (CNCF). All components are available under the Apache 2 License.