From a2237db372c196d3339e66ed05f83a1a64d68fa4 Mon Sep 17 00:00:00 2001
From: Will Baker <wbaker@gmail.com>
Date: Wed, 7 Apr 2021 13:44:43 -0400
Subject: [PATCH 1/2] feat(telemetry): replace unique ids for resources with a
 slug

---
 kit/transport/http/middleware.go      | 56 +++++++++++++++++++++++++++
 kit/transport/http/middleware_test.go | 30 ++++++++++++++
 2 files changed, 86 insertions(+)

diff --git a/kit/transport/http/middleware.go b/kit/transport/http/middleware.go
index 1643736a505..facbc6b8da0 100644
--- a/kit/transport/http/middleware.go
+++ b/kit/transport/http/middleware.go
@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"net/http"
 	"path"
+	"regexp"
 	"strings"
 	"time"
 
@@ -111,17 +112,32 @@ func UserAgent(r *http.Request) string {
 	return ua.Parse(header).Name
 }
 
+// Constants used for normalizing paths
+const (
+	fileSlug  = ":file_name"
+	shardSlug = ":shard_id"
+)
+
 func normalizePath(p string) string {
+	// Normalize any paths used during backup or restore processes
+	p = normalizeBackupAndRestore(p)
+
+	// Go through each part of the path and normalize IDs and UI assets
 	var parts []string
 	for head, tail := shiftPath(p); ; head, tail = shiftPath(tail) {
 		piece := head
+
+		// Normalize any ID's in the path as the ":id" slug
 		if len(piece) == platform.IDLength {
 			if _, err := platform.IDFromString(head); err == nil {
 				piece = ":id"
 			}
 		}
 		parts = append(parts, piece)
+
 		if tail == "/" {
+			// Normalize UI asset file names. The UI asset file is the last part of the path.
+			parts[len(parts)-1] = normalizeAssetFile(parts[len(parts)-1])
 			break
 		}
 	}
@@ -137,6 +153,46 @@ func shiftPath(p string) (head, tail string) {
 	return p[1:i], p[i:]
 }
 
+// Normalize the file name for a UI asset
+// For example: 838442d56d.svg will return as :file_id.svg
+// Files names that do not have one of the listed extensions will be returned unchanged
+func normalizeAssetFile(f string) string {
+	exts := []string{
+		".js",
+		".svg",
+		".woff2",
+		".wasm",
+		".map",
+		".LICENSE",
+	}
+
+	for _, ext := range exts {
+		if strings.HasSuffix(f, ext) {
+			return fileSlug + ext
+		}
+	}
+
+	return f
+}
+
+// Normalize paths used during the backup and restore process.
+// Paths not matching any of the patterns will be returned unchanged.
+func normalizeBackupAndRestore(pth string) string {
+	patterns := map[string]string{
+		`restore/shards/\d+`: path.Join("restore/shards", shardSlug),
+		`backup/shards/\d+`:  path.Join("backup/shards", shardSlug),
+	}
+
+	for p, s := range patterns {
+		re := regexp.MustCompile(p)
+		if re.MatchString(pth) {
+			return re.ReplaceAllString(pth, s)
+		}
+	}
+
+	return pth
+}
+
 type OrgContext string
 
 const CtxOrgKey OrgContext = "orgID"
diff --git a/kit/transport/http/middleware_test.go b/kit/transport/http/middleware_test.go
index 5a500498692..2a6532e09cf 100644
--- a/kit/transport/http/middleware_test.go
+++ b/kit/transport/http/middleware_test.go
@@ -37,6 +37,36 @@ func Test_normalizePath(t *testing.T) {
 			path:     path.Join("/api/v2/organizations", platform.ID(2).String(), "users", platform.ID(3).String()),
 			expected: "/api/v2/organizations/:id/users/:id",
 		},
+		{
+			name:     "5",
+			path:     "/838442d56d.svg",
+			expected: "/" + fileSlug + ".svg",
+		},
+		{
+			name:     "6",
+			path:     "/838442d56d.svg/extra",
+			expected: "/838442d56d.svg/extra",
+		},
+		{
+			name:     "7",
+			path:     "/api/v2/restore/shards/1001",
+			expected: path.Join("/api/v2/restore/shards/", shardSlug),
+		},
+		{
+			name:     "8",
+			path:     "/api/v2/restore/shards/1001/extra",
+			expected: path.Join("/api/v2/restore/shards/", shardSlug, "extra"),
+		},
+		{
+			name:     "9",
+			path:     "/api/v2/backup/shards/1005",
+			expected: path.Join("/api/v2/backup/shards/", shardSlug),
+		},
+		{
+			name:     "10",
+			path:     "/api/v2/backup/shards/1005/extra",
+			expected: path.Join("/api/v2/backup/shards/", shardSlug, "extra"),
+		},
 	}
 
 	for _, tt := range tests {

From ef3aeac4b4b285ea16cb45df88f3e5e709701d8e Mon Sep 17 00:00:00 2001
From: Will Baker <wbaker@gmail.com>
Date: Wed, 7 Apr 2021 15:40:18 -0400
Subject: [PATCH 2/2] chore: update CHANGELOG

---
 CHANGELOG.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 6d4a1f4477c..b15017cf983 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -50,6 +50,7 @@ The prefix used for Prometheus metrics from the query controller has changed fro
 1. [21006](https://github.com/influxdata/influxdb/pull/21006): Add `-p, --profilers` flag to `influx query` command.
 1. [21090](https://github.com/influxdata/influxdb/pull/21090): Update UI to match InfluxDB Cloud.
 1. [21127](https://github.com/influxdata/influxdb/pull/21127): Allow for disabling concurrency-limits in Flux controller.
+1. [21158](https://github.com/influxdata/influxdb/pull/21158): Replace unique resource IDs (UI assets, backup shards) with slugs to reduce cardinality of telemetry data.
 
 ### Bug Fixes