out_influxdb: documentation for tag prefix stripping

pipeline/outputs/influxdb.md

@@ -14,6 +14,7 @@ This plugin supports the following parameters:
 | `Bucket` | InfluxDB bucket name where records will be inserted. If specified, `database` is ignored and v2 of API is used. | _none_ |
 | `Org` | InfluxDB organization name where the bucket is (v2 only). | `fluent` |
 | `Sequence_Tag` | The name of the tag whose value is incremented for the consecutive simultaneous events. | `_seq` |
+| `Strip_Prefix` | String prefix to be removed from the front of `tag` when writing InfluxDB measurement names. `Strip_Prefix` is removed only if it matches exactly at the beginning of and is strictly shorther than `tag` (`tag` length will not be reduced to zero if `Strip_Prefix` and `tag` are identical strings). `Strip_Prefix` defaults to empty string, which does nothing. | _none_ |
 | `HTTP_User` | Optional username for HTTP Basic Authentication. | _none_ |
 | `HTTP_Passwd` | Password for user defined in `HTTP_User`. | _none_ |
 | `HTTP_Token` | Authentication token used with InfluxDB v2. If specified, both `HTTP_User` and `HTTP_Passwd` are ignored. | _none_ |
@@ -191,6 +192,81 @@ pipeline:
 {% endtab %}
 {% endtabs %}
 
+### Prefix stripping
+
+When collecting data from many inputs into many buckets, it can be helpful to remove a common prefix using `Strip_prefix`.
+
+{% tabs %}
+{% tab title="fluent-bit.yaml" %}
+
+```yaml
+pipeline:
+  inputs:
+    - name: cpu
+      tag: cpu.one
+    - name: cpu
+      tag: cpu.two
+    - name: cpu
+      tag: gpu.one
+    - name: cpu
+      tag: gpu.two
+
+
+  outputs:
+    - name: influxdb
+      match: 'cpu*'
+      host: 127.0.0.1
+      port: 8086
+      bucket: cpubucket
+      strip_prefix: cpu.
+    - name: influxdb
+      match: 'gpu*'
+      host: 127.0.0.1
+      port: 8086
+      bucket: gpubucket
+      strip_prefix: gpu.
+```
+
+{% endtab %}
+{% tab title="fluent-bit.conf" %}
+
+```text
+[INPUT]
+    Name  cpu
+    Tag   cpu.one
+
+[INPUT]
+    Name  cpu
+    Tag   cpu.two
+
+[INPUT]
+    Name  cpu
+    Tag   gpu.one
+
+[INPUT]
+    Name  cpu
+    Tag   gpu.two
+
+[OUTPUT]
+    Name          influxdb
+    Match         cpu*
+    Bucket        cpubucket
+    Strip_prefix  cpu.
+
+[OUTPUT]
+    Name          influxdb
+    Match         gpu*
+    Bucket        gpubucket
+    Strip_prefix  gpu.
+```
+
+{% endtab %}
+{% endtabs %}
+
+
+This will result in the buckets `cpubucket` and `gpubucket` each containing two measurement streams named `one` and `two`. Without prefix stripping, the measurement names would be `cpu.one` and `cpu.two` (stored in `cpubucket`), and `gpu.one`, `gpu.two` (stored in `gpubucket`).
+
+
 ### Testing
 
 Before starting Fluent Bit, ensure the target database exists on InfluxDB. Using the previous example, insert the data into a `fluentbit` database.