gitserver: make list of ignored git commands for recording configurable (#55100)

Closes #55059

This PR makes the list of git commands that should be ignored for recording configurable via the site config setting.

## Test plan

<!-- All pull requests REQUIRE a test plan: https://docs.sourcegraph.com/dev/background-information/testing_principles -->
* Add an entry to the `gitRecorder.ignoredGitCommand` array.
* Perform an operation that will execute that command inside of `gitServer`, check redis for the value associated with key `recording-cmd` and it shouldn't be part of the result.

cmd/gitserver/shared/BUILD.bazel

@@ -18,6 +18,7 @@ go_library(
         "//internal/api",
         "//internal/authz",
         "//internal/codeintel/dependencies",
+        "//internal/collections",
         "//internal/conf",
         "//internal/conf/conftypes",
         "//internal/database",

cmd/gitserver/shared/shared.go

@@ -28,6 +28,7 @@ import (
 	"github.com/sourcegraph/sourcegraph/internal/api"
 	"github.com/sourcegraph/sourcegraph/internal/authz"
 	"github.com/sourcegraph/sourcegraph/internal/codeintel/dependencies"
+	"github.com/sourcegraph/sourcegraph/internal/collections"
 	"github.com/sourcegraph/sourcegraph/internal/conf"
 	"github.com/sourcegraph/sourcegraph/internal/conf/conftypes"
 	"github.com/sourcegraph/sourcegraph/internal/database"
@@ -146,7 +147,7 @@ func Main(ctx context.Context, observationCtx *observation.Context, ready servic
 			recordingCommandFactory.Disable()
 			return
 		}
-		recordingCommandFactory.Update(recordCommandsOnRepos(recordingConf.Repos), recordingConf.Size)
+		recordingCommandFactory.Update(recordCommandsOnRepos(recordingConf.Repos, recordingConf.IgnoredGitCommands), recordingConf.Size)
 	})
 
 	gitserver := server.Server{
@@ -615,9 +616,17 @@ func methodSpecificUnaryInterceptor(method string, next grpc.UnaryServerIntercep
 	}
 }
 
+var defaultIgnoredGitCommands = []string{
+	"show",
+	"rev-parse",
+	"log",
+	"diff",
+	"ls-tree",
+}
+
 // recordCommandsOnRepos returns a ShouldRecordFunc which determines whether the given command should be recorded
 // for a particular repository.
-func recordCommandsOnRepos(repos []string) wrexec.ShouldRecordFunc {
+func recordCommandsOnRepos(repos []string, ignoredGitCommands []string) wrexec.ShouldRecordFunc {
 	// empty repos, means we should never record since there is nothing to match on
 	if len(repos) == 0 {
 		return func(ctx context.Context, c *exec.Cmd) bool {
@@ -625,14 +634,13 @@ func recordCommandsOnRepos(repos []string) wrexec.ShouldRecordFunc {
 		}
 	}
 
-	// we won't record any git commands with these commands since they are considered to be not destructive
-	ignoredGitCommands := map[string]struct{}{
-		"show":      {},
-		"rev-parse": {},
-		"log":       {},
-		"diff":      {},
-		"ls-tree":   {},
+	if len(ignoredGitCommands) == 0 {
+		ignoredGitCommands = append(ignoredGitCommands, defaultIgnoredGitCommands...)
 	}
+
+	// we won't record any git commands with these commands since they are considered to be not destructive
+	var ignoredGitCommandsMap = collections.NewSet(ignoredGitCommands...)
+
 	return func(ctx context.Context, cmd *exec.Cmd) bool {
 		base := filepath.Base(cmd.Path)
 		if base != "git" {
@@ -658,7 +666,7 @@ func recordCommandsOnRepos(repos []string) wrexec.ShouldRecordFunc {
 		// we have to scan the Args, since it isn't guaranteed that the Arg at index 1 is the git command:
 		// git -c "protocol.version=2" remote show
 		for _, arg := range cmd.Args {
-			if _, ok := ignoredGitCommands[arg]; ok {
+			if ok := ignoredGitCommandsMap.Has(arg); ok {
 				return false
 			}
 		}

schema/schema.go

@@ -1311,6 +1311,8 @@ type GitLabWebhook struct {
 
 // GitRecorder description: Record git operations that are executed on configured repositories. The following commands are not recorded: show, log, rev-parse and diff.
 type GitRecorder struct {
+	// IgnoredGitCommands description: List of git commands that should be ignored and not recorded.
+	IgnoredGitCommands []string `json:"ignoredGitCommands,omitempty"`
 	// Repos description: List of repositories whose git operations should be recorded.
 	Repos []string `json:"repos,omitempty"`
 	// Size description: Defines how many recordings to keep. Once this size is reached, the oldest entry will be removed.

schema/site.schema.json

@@ -1905,7 +1905,9 @@
         "size": {
           "description": "Defines how many recordings to keep. Once this size is reached, the oldest entry will be removed.",
           "type": "integer",
-          "default": 100000
+          "default": 10000,
+          "minimum": 1,
+          "maximum": 10000
         },
         "repos": {
           "description": "List of repositories whose git operations should be recorded.",
@@ -1913,12 +1915,21 @@
           "items": {
             "type": "string"
           }
+        },
+        "ignoredGitCommands": {
+          "description": "List of git commands that should be ignored and not recorded.",
+          "type": "array",
+          "items": {
+            "type": "string"
+          },
+          "default": ["show", "rev-parse", "log", "diff", "ls-tree"]
         }
       },
       "examples": [
         {
           "size": 1000,
-          "repos": ["github.com/sourcegraph/sourcegraph", "github.com/gorilla/mux"]
+          "repos": ["github.com/sourcegraph/sourcegraph", "github.com/gorilla/mux"],
+          "ignoredGitCommands": ["show", "rev-parse", "log", "diff", "ls-tree"]
         }
       ]
     },