fix: allow promtail docker_sd_config to handle long log lines (#6256)

In promtail's docker_sd_config, parsing of a container's
logs would fail if any log line was over 64k bytes in length.
This is because bufio.Scanner uses a 64k buffer by default.
Since there doesn't appear to be a way to elegantly catch
`bufio.ErrTooLong` and carry on, the parsing approach was changed
to bufio.Reader which can handle lines of arbitrary length.

clients/pkg/promtail/targets/docker/target.go

@@ -148,14 +148,38 @@ func extractTs(line string) (time.Time, string, error) {
 	return ts, pair[1], nil
 }
 
+// https://devmarkpro.com/working-big-files-golang
+func readLine(r *bufio.Reader) (string, error) {
+	var (
+		isPrefix = true
+		err      error
+		line, ln []byte
+	)
+
+	for isPrefix && err == nil {
+		line, isPrefix, err = r.ReadLine()
+		ln = append(ln, line...)
+	}
+
+	return string(ln), err
+}
+
 func (t *Target) process(r io.Reader, logStream string) {
 	defer func() {
 		t.wg.Done()
 	}()
 
-	scanner := bufio.NewScanner(r)
-	for scanner.Scan() {
-		line := scanner.Text()
+	reader := bufio.NewReader(r)
+	for {
+		line, err := readLine(reader)
+		if err != nil {
+			if err == io.EOF {
+				break
+			}
+			level.Error(t.logger).Log("msg", "error reading docker log line, skipping line", "err", err)
+			t.metrics.dockerErrors.Inc()
+		}
+
 		ts, line, err := extractTs(line)
 		if err != nil {
 			level.Error(t.logger).Log("msg", "could not extract timestamp, skipping line", "err", err)
@@ -189,12 +213,6 @@ func (t *Target) process(r io.Reader, logStream string) {
 		t.metrics.dockerEntries.Inc()
 		t.positions.Put(positions.CursorKey(t.containerName), ts.Unix())
 	}
-
-	err := scanner.Err()
-	if err != nil {
-		level.Warn(t.logger).Log("msg", "finished scanning logs lines with an error", "err", err)
-	}
-
 }
 
 // startIfNotRunning starts processing container logs. The operation is idempotent , i.e. the processing cannot be started twice.

clients/pkg/promtail/targets/docker/targetmanager_test.go

@@ -89,7 +89,7 @@ func Test_TargetManager(t *testing.T) {
 	require.True(t, ta.Ready())
 
 	require.Eventually(t, func() bool {
-		return len(entryHandler.Received()) >= 5
+		return len(entryHandler.Received()) >= 6
 	}, 20*time.Second, 100*time.Millisecond)
 
 	received := entryHandler.Received()
@@ -109,4 +109,5 @@ func Test_TargetManager(t *testing.T) {
 		actualLines = append(actualLines, entry.Line)
 	}
 	require.ElementsMatch(t, actualLines, expectedLines)
+	require.Equal(t, 99969, len(received[5].Line))
 }