From e389b7b9c49ae7d53f44864655848eb225dd7aa9 Mon Sep 17 00:00:00 2001
From: Joshua Powers <powersj@fastmail.com>
Date: Fri, 10 May 2024 10:25:51 -0600
Subject: [PATCH] feat(inputs.file): Add tag with absolute path of file
 (#15330)

---
 plugins/inputs/file/README.md    | 13 +++++++++----
 plugins/inputs/file/file.go      |  6 ++++++
 plugins/inputs/file/file_test.go | 14 ++++++++------
 plugins/inputs/file/sample.conf  | 13 +++++++++----
 4 files changed, 32 insertions(+), 14 deletions(-)

diff --git a/plugins/inputs/file/README.md b/plugins/inputs/file/README.md
index f732125aefefe..8870c69d3a21e 100644
--- a/plugins/inputs/file/README.md
+++ b/plugins/inputs/file/README.md
@@ -39,12 +39,17 @@ See the [CONFIGURATION.md][CONFIGURATION.md] for more details.
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "influx"
 
-
-  ## Name a tag containing the name of the file the data was parsed from.  Leave empty
-  ## to disable. Cautious when file name variation is high, this can increase the cardinality
-  ## significantly. Read more about cardinality here:
+  ## Please use caution when using the following options: when file name
+  ## variation is high, this can increase the cardinality significantly. Read
+  ## more about cardinality here:
   ## https://docs.influxdata.com/influxdb/cloud/reference/glossary/#series-cardinality
+
+  ## Name of tag to store the name of the file. Disabled if not set.
   # file_tag = ""
+
+  ## Name of tag to store the absolute path and name of the file. Disabled if
+  ## not set.
+  # file_path_tag = ""
 ```
 
 ## Metrics
diff --git a/plugins/inputs/file/file.go b/plugins/inputs/file/file.go
index 84d989fb0bd36..a0b13c472e85e 100644
--- a/plugins/inputs/file/file.go
+++ b/plugins/inputs/file/file.go
@@ -22,6 +22,7 @@ var sampleConfig string
 type File struct {
 	Files             []string `toml:"files"`
 	FileTag           string   `toml:"file_tag"`
+	FilePathTag       string   `toml:"file_path_tag"`
 	CharacterEncoding string   `toml:"character_encoding"`
 
 	parserFunc telegraf.ParserFunc
@@ -54,6 +55,11 @@ func (f *File) Gather(acc telegraf.Accumulator) error {
 			if f.FileTag != "" {
 				m.AddTag(f.FileTag, filepath.Base(k))
 			}
+			if f.FilePathTag != "" {
+				if absPath, err := filepath.Abs(k); err == nil {
+					m.AddTag(f.FilePathTag, absPath)
+				}
+			}
 			acc.AddMetric(m)
 		}
 	}
diff --git a/plugins/inputs/file/file_test.go b/plugins/inputs/file/file_test.go
index cd539bd5a5fe0..91e7fc8a8a134 100644
--- a/plugins/inputs/file/file_test.go
+++ b/plugins/inputs/file/file_test.go
@@ -42,8 +42,9 @@ func TestFileTag(t *testing.T) {
 	wd, err := os.Getwd()
 	require.NoError(t, err)
 	r := File{
-		Files:   []string{filepath.Join(wd, "dev", "testfiles", "json_a.log")},
-		FileTag: "filename",
+		Files:       []string{filepath.Join(wd, "dev", "testfiles", "json_a.log")},
+		FileTag:     "filename",
+		FilePathTag: "filepath",
 	}
 	require.NoError(t, r.Init())
 
@@ -56,10 +57,11 @@ func TestFileTag(t *testing.T) {
 	require.NoError(t, r.Gather(&acc))
 
 	for _, m := range acc.Metrics {
-		for key, value := range m.Tags {
-			require.Equal(t, r.FileTag, key)
-			require.Equal(t, filepath.Base(r.Files[0]), value)
-		}
+		require.Contains(t, m.Tags, "filename")
+		require.Equal(t, filepath.Base(r.Files[0]), m.Tags["filename"])
+
+		require.Contains(t, m.Tags, "filepath")
+		require.True(t, filepath.IsAbs(m.Tags["filepath"]))
 	}
 }
 
diff --git a/plugins/inputs/file/sample.conf b/plugins/inputs/file/sample.conf
index 706d1bfff6481..9f8788d4b6388 100644
--- a/plugins/inputs/file/sample.conf
+++ b/plugins/inputs/file/sample.conf
@@ -19,9 +19,14 @@
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "influx"
 
-
-  ## Name a tag containing the name of the file the data was parsed from.  Leave empty
-  ## to disable. Cautious when file name variation is high, this can increase the cardinality
-  ## significantly. Read more about cardinality here:
+  ## Please use caution when using the following options: when file name
+  ## variation is high, this can increase the cardinality significantly. Read
+  ## more about cardinality here:
   ## https://docs.influxdata.com/influxdb/cloud/reference/glossary/#series-cardinality
+
+  ## Name of tag to store the name of the file. Disabled if not set.
   # file_tag = ""
+
+  ## Name of tag to store the absolute path and name of the file. Disabled if
+  ## not set.
+  # file_path_tag = ""