fix: sanitize workspace agent logs before insert (#24028)

Workspace agent logs could still fail after the earlier invalid UTF-8
fix because NUL bytes are valid Go/protobuf strings but are rejected by
Postgres text columns. The legacy HTTP log upload path also bypassed the
old sanitization entirely, and both server insert paths computed
logs_length from the unsanitized input.

Add a shared log-output sanitizer in agentsdk, use it in the protobuf
conversion path and both server-side insert paths, and compute
OutputLength from the sanitized string so overflow accounting matches
what is actually stored. This keeps the old invalid UTF-8 behavior while
also handling embedded NUL bytes consistently across DRPC and HTTP log
ingestion.

Refs [#23292 ](https://github.com/coder/coder/issues/23292)
Refs [#13433 ](https://github.com/coder/coder/issues/13433)

codersdk/agentsdk/convert.go

@@ -376,7 +376,7 @@ func ProtoFromLog(log Log) (*proto.Log, error) {
 	}
 	return &proto.Log{
 		CreatedAt: timestamppb.New(log.CreatedAt),
-		Output:    strings.ToValidUTF8(log.Output, "❌"),
+		Output:    SanitizeLogOutput(log.Output),
 		Level:     proto.Log_Level(lvl),
 	}, nil
 }

codersdk/agentsdk/logs_internal_test.go

@@ -229,7 +229,7 @@ func TestLogSender_SkipHugeLog(t *testing.T) {
 	require.ErrorIs(t, err, context.Canceled)
 }
 
-func TestLogSender_InvalidUTF8(t *testing.T) {
+func TestLogSender_SanitizeOutput(t *testing.T) {
 	t.Parallel()
 	testCtx := testutil.Context(t, testutil.WaitShort)
 	ctx, cancel := context.WithCancel(testCtx)
@@ -243,7 +243,7 @@ func TestLogSender_InvalidUTF8(t *testing.T) {
 	uut.Enqueue(ls1,
 		Log{
 			CreatedAt: t0,
-			Output:    "test log 0, src 1\xc3\x28",
+			Output:    "test log 0, src 1\x00\xc3\x28",
 			Level:     codersdk.LogLevelInfo,
 		},
 		Log{
@@ -260,10 +260,10 @@ func TestLogSender_InvalidUTF8(t *testing.T) {
 
 	req := testutil.TryReceive(ctx, t, fDest.reqs)
 	require.NotNil(t, req)
-	require.Len(t, req.Logs, 2, "it should sanitize invalid UTF-8, but still send")
-	// the 0xc3, 0x28 is an invalid 2-byte sequence in UTF-8.  The sanitizer replaces 0xc3 with ❌, and then
-	// interprets 0x28 as a 1-byte sequence "("
-	require.Equal(t, "test log 0, src 1❌(", req.Logs[0].GetOutput())
+	require.Len(t, req.Logs, 2, "it should sanitize invalid output, but still send")
+	// The sanitizer replaces the NUL byte and invalid UTF-8 with ❌ while
+	// preserving the valid "(" byte that follows 0xc3.
+	require.Equal(t, "test log 0, src 1❌❌(", req.Logs[0].GetOutput())
 	require.Equal(t, proto.Log_INFO, req.Logs[0].GetLevel())
 	require.Equal(t, "test log 1, src 1", req.Logs[1].GetOutput())
 	require.Equal(t, proto.Log_INFO, req.Logs[1].GetLevel())

codersdk/agentsdk/logs_sanitize.go

@@ -0,0 +1,11 @@
+package agentsdk
+
+import "strings"
+
+// SanitizeLogOutput replaces invalid UTF-8 and NUL characters in log output.
+// Invalid UTF-8 cannot be transported in protobuf string fields, and PostgreSQL
+// rejects NUL bytes in text columns.
+func SanitizeLogOutput(s string) string {
+	s = strings.ToValidUTF8(s, "❌")
+	return strings.ReplaceAll(s, "\x00", "❌")
+}

codersdk/agentsdk/logs_test.go

@@ -17,6 +17,54 @@ import (
 	"github.com/coder/coder/v2/testutil"
 )
 
+func TestSanitizeLogOutput(t *testing.T) {
+	t.Parallel()
+
+	tests := []struct {
+		name string
+		in   string
+		want string
+	}{
+		{
+			name: "valid",
+			in:   "hello world",
+			want: "hello world",
+		},
+		{
+			name: "invalid utf8",
+			in:   "test log\xc3\x28",
+			want: "test log❌(",
+		},
+		{
+			name: "nul byte",
+			in:   "before\x00after",
+			want: "before❌after",
+		},
+		{
+			name: "invalid utf8 and nul byte",
+			in:   "before\x00middle\xc3\x28after",
+			want: "before❌middle❌(after",
+		},
+		{
+			name: "nul byte at edges",
+			in:   "\x00middle\x00",
+			want: "❌middle❌",
+		},
+		{
+			name: "invalid utf8 at edges",
+			in:   "\xc3middle\xc3",
+			want: "❌middle❌",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+			require.Equal(t, tt.want, agentsdk.SanitizeLogOutput(tt.in))
+		})
+	}
+}
+
 func TestStartupLogsWriter_Write(t *testing.T) {
 	t.Parallel()