Issue 2864 - Introduce class logstash

modules/logstash/manifests/pipeline.pp

Index: modules/logstash/manifests/pipeline.pp
diff --git a/modules/logstash/manifests/pipeline.pp b/modules/logstash/manifests/pipeline.pp
new file mode 100644
index 0000000000000000000000000000000000000000..7d16cee2a528960e63c79b68bcdaa3c27be29371
--- /dev/null
+++ b/modules/logstash/manifests/pipeline.pp
@@ -0,0 +1,87 @@
+# == Type: logstash::pipeline
+#
+# Manage Logstash (https://logstash.net/) pipeline configuration files.
+#
+# While one can directly provide the configuration $source or $content, one
+# should note that a logstash::pipeline resource is actually the head of the
+# logfile for concatenation (analogous to file_concat or module concat). Use
+# logstash::fragment to assemble such a file from multiple resources.
+#
+# === Parameters:
+#
+# [*content*]
+#   The configuration as-is, mutually exclusive with $source.
+#
+# [*description*]
+#   If both $content and $source are undef, $description is converted into
+#   a comment header (prefixed by #) for the pipeline configuration, which
+#   is then expected to be setup further via logstash::fragment.
+#
+# [*ensure*]
+#   Either 'present' or 'absent'.
+#
+# [*source*]
+#   The configuration source location, mutually exclusive with $content.
+#
+# === Examples:
+#
+# Below please find a list of examples on how to use this type. Examples
+# on howto configure a Logstash pipeline are provided by Elastic at

[Felix Dahlke, 2015/10/07 10:46:32]

Nit: "how to"

[mathias, 2015/10/20 13:03:34]

Done.

+# https://www.elastic.co/guide/en/logstash/current/config-examples.html
+#
+#   logstash::pipeline {'alpha':
+#     description => 'A pipeline setup using logstash::fragment',
+#     ensure => 'present',
+#   }
+#
+#   logstash::pipeline {'beta':
+#     description => 'A pipeline setup from a single template',
+#     content => template('custom/pipeline.erb'),
+#   }
+#
+#   logstash::pipeline {'gamma':
+#     description => 'An obsolete setup to be removed if present',
+#     ensure => 'absent',
+#   }
+#
+# For more information on howto use logstash::fragment with a pipeline

[Felix Dahlke, 2015/10/07 10:46:32]

Nit: "how to"

[mathias, 2015/10/20 13:03:34]

Done.

+# like 'alpha' above can be found in the accompanying fragment.pp file.
+#
+define logstash::pipeline(
+  $content = undef,
+  $description = "Puppet: Logstash::Pipeline['$title']",
+  $ensure = 'present',
+  $source = undef,
+) {
+
+  $id = "logstash::pipeline#$title"
+  $path = sprintf("/etc/logstash/conf.d/puppet-%s.conf", uriescape($title))
+
+  if $ensure !~ /^(absent|purged)$/ {
+
+    concat {$id:
+      notify => Service['logstash'],
+      path => $path,
+      require => Package['logstash'],
+    }
+
+    concat::fragment {$id:
+      content => $content ? {
+        undef => $source ? {
+          undef => sprintf("# %s\n\n", join("\n# ", split($description, "\n"))),

[Felix Dahlke, 2015/10/07 10:46:33]

Why do we add anything if there's neither content or source? Do we need to add
_something_?

[mathias, 2015/10/20 13:03:34]

One must supply either $content or $source for concat::fragment, unless $ensure
is "absent" that is.

This particular approach here was meant as an easy way to setup a header for a
pipeline configuration that is spread over multiple fragments. When introduced,
logstash::pipeline did not feature $content and $source yet. However, in it's
current form this now seems rather obsolete. One can achieve pretty much the
same by using $content. I'll remove it.

+          default => undef,
+        },
+        default => $content,
+      },
+      order => 0,
+      source => $source,
+      target => $id,
+    }
+  }
+  elsif !defined(File[$path]) {
+
+    file {$path:
+      ensure => 'absent',
+    }
+  }
+}