Optimize queries using regex matchers for set lookups

labels/selector.go

@@ -63,14 +63,15 @@ func NewEqualMatcher(name, value string) Matcher {
 	return &EqualMatcher{name: name, value: value}
 }
 
-type regexpMatcher struct {
+type RegexpMatcher struct {
 	name string
 	re   *regexp.Regexp
 }
 
-func (m regexpMatcher) Name() string          { return m.name }
-func (m regexpMatcher) Matches(v string) bool { return m.re.MatchString(v) }
-func (m regexpMatcher) String() string        { return fmt.Sprintf("%s=~%q", m.name, m.re.String()) }
+func (m RegexpMatcher) Name() string          { return m.name }
+func (m RegexpMatcher) Matches(v string) bool { return m.re.MatchString(v) }
+func (m RegexpMatcher) String() string        { return fmt.Sprintf("%s=~%q", m.name, m.re.String()) }
+func (m RegexpMatcher) Value() string         { return m.re.String() }
 
 // NewRegexpMatcher returns a new matcher verifying that a value matches
 // the regular expression pattern.
@@ -79,7 +80,7 @@ func NewRegexpMatcher(name, pattern string) (Matcher, error) {
 	if err != nil {
 		return nil, err
 	}
-	return &regexpMatcher{name: name, re: re}, nil
+	return &RegexpMatcher{name: name, re: re}, nil
 }
 
 // NewMustRegexpMatcher returns a new matcher verifying that a value matches
@@ -90,7 +91,7 @@ func NewMustRegexpMatcher(name, pattern string) Matcher {
 	if err != nil {
 		panic(err)
 	}
-	return &regexpMatcher{name: name, re: re}
+	return &RegexpMatcher{name: name, re: re}
 
 }
 

querier.go

@@ -17,6 +17,7 @@ import (
 	"fmt"
 	"sort"
 	"strings"
+	"unicode/utf8"
 
 	"github.com/pkg/errors"
 	"github.com/prometheus/tsdb/chunkenc"
@@ -266,6 +267,62 @@ func (q *blockQuerier) Close() error {
 	return merr.Err()
 }
 
+// Bitmap used by func special to check whether a character needs to be escaped.
+var specialBytes [16]byte
+
+// special reports whether byte b needs to be escaped.
+func special(b byte) bool {
+	return b < utf8.RuneSelf && specialBytes[b%16]&(1<<(b/16)) != 0
+}
+
+func init() {
+	for _, b := range []byte(`.+*?()|[]{}^$`) {
+		specialBytes[b%16] |= 1 << (b / 16)
+	}
+}
+
+func findSetMatches(pattern string) []string {
+	// Return empty matches if the wrapper from Prometheus is missing.
+	if len(pattern) < 6 || pattern[:4] != "^(?:" || pattern[len(pattern)-2:] != ")$" {
+		return []string{}
+	}
+	escaped := false
+	sets := []*strings.Builder{&strings.Builder{}}
+	for i := 4; i < len(pattern)-2; i++ {
+		if escaped {
+			// Add the escaped special character to the sets.
+			if special(pattern[i]) {
+				sets[len(sets)-1].WriteByte(pattern[i])
+			} else if pattern[i] == '\\' {
+				sets[len(sets)-1].WriteByte('\\')
+			} else {
+				return []string{}
+			}
+			escaped = false
+		} else {
+			// Return empty sets when there are special characters excluding '|'.
+			if special(pattern[i]) {
+				if pattern[i] == '|' {
+					sets = append(sets, &strings.Builder{})
+				} else {
+					return []string{}
+				}
+			} else if pattern[i] == '\\' {
+				escaped = true
+			} else {
+				sets[len(sets)-1].WriteByte(pattern[i])
+			}
+		}
+	}
+	matches := make([]string, 0, len(sets))
+	for _, s := range sets {
+		if s.Len() > 0 {
+			matches = append(matches, s.String())
+		}
+	}
+	return matches
+}
+
 // PostingsForMatchers assembles a single postings iterator against the index reader
 // based on the given matchers.
 func PostingsForMatchers(ix IndexReader, ms ...labels.Matcher) (index.Postings, error) {
@@ -346,6 +403,14 @@ func postingsForMatcher(ix IndexReader, m labels.Matcher) (index.Postings, error
 		return ix.Postings(em.Name(), em.Value())
 	}
 
+	// Fast-path for set matching.
+	if em, ok := m.(*labels.RegexpMatcher); ok {
+		setMatches := findSetMatches(em.Value())
+		if len(setMatches) > 0 {
+			return postingsForSetMatcher(ix, em.Name(), setMatches)
+		}
+	}
+
 	tpls, err := ix.LabelValues(m.Name())
 	if err != nil {
 		return nil, err
@@ -411,6 +476,16 @@ func inversePostingsForMatcher(ix IndexReader, m labels.Matcher) (index.Postings
 	return index.Merge(rit...), nil
 }
 
+func postingsForSetMatcher(ix IndexReader, name string, matches []string) (index.Postings, error) {
+	var its []index.Postings
+	for _, match := range matches {
+		if it, err := ix.Postings(name, match); err == nil {
+			its = append(its, it)
+		}
+	}
+	return index.Merge(its...), nil
+}
+
 func mergeStrings(a, b []string) []string {
 	maxl := len(a)
 	if len(b) > len(a) {