Collector implementation

cmd/collector/parser.go

@@ -0,0 +1,71 @@
+package main
+
+import (
+	"fmt"
+	"net"
+	"strings"
+)
+
+// LogRecord holds the four dimensions extracted from a single nginx log line.
+type LogRecord struct {
+	Website      string
+	ClientPrefix string
+	URI          string
+	Status       string
+}
+
+// ParseLine parses a tab-separated logtail log line:
+//
+//	$host \t $remote_addr \t $msec \t $request_method \t $request_uri \t $status \t $body_bytes_sent \t $request_time
+//
+// Returns false for lines with fewer than 8 fields.
+func ParseLine(line string, v4bits, v6bits int) (LogRecord, bool) {
+	// SplitN caps allocations; we need exactly 8 fields.
+	fields := strings.SplitN(line, "\t", 8)
+	if len(fields) < 8 {
+		return LogRecord{}, false
+	}
+
+	uri := fields[4]
+	if i := strings.IndexByte(uri, '?'); i >= 0 {
+		uri = uri[:i]
+	}
+
+	prefix, ok := truncateIP(fields[1], v4bits, v6bits)
+	if !ok {
+		return LogRecord{}, false
+	}
+
+	return LogRecord{
+		Website:      fields[0],
+		ClientPrefix: prefix,
+		URI:          uri,
+		Status:       fields[5],
+	}, true
+}
+
+// truncateIP masks addr to the given prefix length depending on IP version.
+// Returns the CIDR string (e.g. "1.2.3.0/24") and true on success.
+func truncateIP(addr string, v4bits, v6bits int) (string, bool) {
+	ip := net.ParseIP(addr)
+	if ip == nil {
+		return "", false
+	}
+
+	var bits int
+	if ip.To4() != nil {
+		ip = ip.To4()
+		bits = v4bits
+	} else {
+		ip = ip.To16()
+		bits = v6bits
+	}
+
+	mask := net.CIDRMask(bits, len(ip)*8)
+	masked := make(net.IP, len(ip))
+	for i := range ip {
+		masked[i] = ip[i] & mask[i]
+	}
+
+	return fmt.Sprintf("%s/%d", masked.String(), bits), true
+}