[SQL Storage indexer] Add cache for categories endpoint (#1379)

This PR adds an independent cache for the categories endpoint
(/categories) following the same pattern as it was implemented
for the search endpoint.

Author: mrodm

CHANGELOG.md

@@ -14,6 +14,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 * New Security subcategory "asset_inventory" [#1357](https://github.com/elastic/package-registry/pull/1357)
 * Update default value for the batch size used in SQL storage indexer. [#1372](https://github.com/elastic/package-registry/pull/1372)
 * Skip adding to cache requests containing package query parameter. [#1378](https://github.com/elastic/package-registry/pull/1378)
+* Added cache for categories endpoint (SQL storage indexer). [#1379](https://github.com/elastic/package-registry/1379)
+* Update default values for cache search and TTL settings search (SQL storage indexer). [#1379](https://github.com/elastic/package-registry/1379)
 
 ### Deprecated
 

categories.go

@@ -18,6 +18,7 @@ import (
 	"go.uber.org/zap"
 
 	"github.com/Masterminds/semver/v3"
+	"github.com/hashicorp/golang-lru/v2/expirable"
 	"go.elastic.co/apm/module/apmzap/v2"
 	"go.elastic.co/apm/v2"
 
@@ -28,14 +29,22 @@ import (
 
 // categoriesHandler is a dynamic handler as it will also allow filtering in the future.
 func categoriesHandler(logger *zap.Logger, indexer Indexer, cacheTime time.Duration) func(w http.ResponseWriter, r *http.Request) {
-	return categoriesHandlerWithProxyMode(logger, indexer, proxymode.NoProxy(logger), cacheTime)
+	return categoriesHandlerWithProxyMode(logger, indexer, proxymode.NoProxy(logger), cacheTime, nil)
 }
 
 // categoriesHandler is a dynamic handler as it will also allow filtering in the future.
-func categoriesHandlerWithProxyMode(logger *zap.Logger, indexer Indexer, proxyMode *proxymode.ProxyMode, cacheTime time.Duration) func(w http.ResponseWriter, r *http.Request) {
+func categoriesHandlerWithProxyMode(logger *zap.Logger, indexer Indexer, proxyMode *proxymode.ProxyMode, cacheTime time.Duration, cache *expirable.LRU[string, []byte]) func(w http.ResponseWriter, r *http.Request) {
 	return func(w http.ResponseWriter, r *http.Request) {
 		logger := logger.With(apmzap.TraceContext(r.Context())...)
 
+		if cache != nil {
+			if response, ok := cache.Get(r.URL.String()); ok {
+				logger.Debug("using as response cached request", zap.String("cache.url", r.URL.String()), zap.Int("cache.size", cache.Len()))
+				serveJSONResponse(r.Context(), w, cacheTime, response)
+				return
+			}
+		}
+
 		query := r.URL.Query()
 
 		filter, err := newCategoriesFilterFromQuery(query)
@@ -93,6 +102,11 @@ func categoriesHandlerWithProxyMode(logger *zap.Logger, indexer Indexer, proxyMo
 		}
 
 		serveJSONResponse(r.Context(), w, cacheTime, data)
+
+		if cache != nil {
+			val := cache.Add(r.URL.String(), data)
+			logger.Debug("added to cache request", zap.String("cache.url", r.URL.String()), zap.Int("cache.size", cache.Len()), zap.Bool("cache.eviction", val))
+		}
 	}
 }
 

categories_test.go

@@ -54,7 +54,7 @@ func TestCategoriesWithProxyMode(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	categoriesWithProxyHandler := categoriesHandlerWithProxyMode(testLogger, indexerProxy, proxyMode, testCacheTime)
+	categoriesWithProxyHandler := categoriesHandlerWithProxyMode(testLogger, indexerProxy, proxyMode, testCacheTime, nil)
 
 	tests := []struct {
 		endpoint string

dev/launch_epr_service_storage_indexer.sh

@@ -19,7 +19,7 @@ usage() {
     echo -e "\t\t\tIf set, the bucket name will be ignored (-b parameter) and Package Registry will use its default development bucket gs://fake-package-storage-internal"
     echo -e "\t-c <config_path>: Path to the configurastion file. Default: \"\""
     echo -e "\t-s : Enable SQL Storage indexer. By default Storage Indexer is enabled."
-    echo -e "\t-C : Enable Search Cache. Just supported with SQL Storage indexer. By default Search Cache is disabled."
+    echo -e "\t-C : Enable Cache for Search and Categories endpoints. Just supported with the SQL Storage indexer. By default both Caches are disabled."
     echo -e "\t-d: Enable debug mode. Default: false"
     echo -e "\t-h: Show this message"
 }
@@ -30,8 +30,7 @@ INDEX_PATH=""
 EMULATOR_HOST="localhost:4443"
 CONFIG_PATH="${SCRIPT_DIR}/../config.yml"
 ENABLE_STORAGE_SQL_INDEXER=0
-ENABLE_SEARCH_CACHE=0
-ENABLE_DEBUG_MODE=0
+ENABLE_CACHE=0
 
 while getopts ":b:p:i:e:c:sdCh" o; do
   case "${o}" in
@@ -54,7 +53,7 @@ while getopts ":b:p:i:e:c:sdCh" o; do
       ENABLE_STORAGE_SQL_INDEXER=1
       ;;
     C)
-      ENABLE_SEARCH_CACHE=1
+      ENABLE_CACHE=1
       ;;
     d)
       ENABLE_DEBUG_MODE=1
@@ -101,8 +100,9 @@ else
     export EPR_FEATURE_STORAGE_INDEXER="false"
 fi
 
-if [[ "${ENABLE_SEARCH_CACHE}" == 1 ]]; then
+if [[ "${ENABLE_CACHE}" == 1 ]]; then
     export EPR_FEATURE_ENABLE_SEARCH_CACHE="true"
+    export EPR_FEATURE_ENABLE_CATEGORIES_CACHE="true"
 fi
 
 export EPR_DISABLE_PACKAGE_VALIDATION="true"
@@ -113,8 +113,10 @@ if [[ "${ENABLE_DEBUG_MODE}" == 1 ]]; then
 fi
 export EPR_CONFIG="${CONFIG_PATH}"
 # export EPR_SQL_INDEXER_DATABASE_FOLDER_PATH=/tmp
-# export EPR_SQL_INDEXER_SEARCH_CACHE_SIZE=100
-# export EPR_SQL_INDEXER_SEARCH_CACHE_TTL=10m
+# export EPR_SQL_INDEXER_SEARCH_CACHE_SIZE=250
+# export EPR_SQL_INDEXER_SEARCH_CACHE_TTL=24h
+# export EPR_SQL_INDEXER_CATEGORIES_CACHE_SIZE=100
+# export EPR_SQL_INDEXER_CATEGORIES_CACHE_TTL=24h
 # export EPR_SQL_INDEXER_READ_PACKAGES_BATCH_SIZE=2000
 # export EPR_SQL_INDEXER_DB_INSERT_BATCH_SIZE=2000
 

internal/storage/sqlindexer.go

@@ -16,8 +16,6 @@ import (
 	"sync"
 	"time"
 
-	"github.com/hashicorp/golang-lru/v2/expirable"
-
 	"cloud.google.com/go/storage"
 
 	"github.com/prometheus/client_golang/prometheus"
@@ -59,7 +57,7 @@ type SQLIndexer struct {
 
 	readPackagesBatchSize int
 
-	cache *expirable.LRU[string, []byte] // Cache for search results
+	afterUpdateHook func(ctx context.Context)
 }
 
 type IndexerOptions struct {
@@ -69,8 +67,8 @@ type IndexerOptions struct {
 	WatchInterval                time.Duration
 	Database                     database.Repository
 	SwapDatabase                 database.Repository
-	Cache                        *expirable.LRU[string, []byte] // Cache for search results
 	ReadPackagesBatchsize        int
+	AfterUpdateIndexHook         func(ctx context.Context)
 }
 
 func NewIndexer(logger *zap.Logger, storageClient *storage.Client, options IndexerOptions) *SQLIndexer {
@@ -86,8 +84,8 @@ func NewIndexer(logger *zap.Logger, storageClient *storage.Client, options Index
 		swapDatabase:          options.SwapDatabase,
 		label:                 fmt.Sprintf("storage-%s", options.PackageStorageEndpoint),
 		readPackagesBatchSize: defaultReadPackagesBatchSize,
-		cache:                 options.Cache,
 		cursor:                "init",
+		afterUpdateHook:       options.AfterUpdateIndexHook,
 	}
 
 	indexer.current = &indexer.database
@@ -302,10 +300,8 @@ func (i *SQLIndexer) swapDatabases(ctx context.Context, currentCursor string, nu
 	i.current, i.backup = i.backup, i.current
 	i.logger.Debug("Current database changed", zap.String("current.database.path", (*i.current).File(ctx)), zap.String("previous.database.path", (*i.backup).File(ctx)))
 
-	if i.cache != nil {
-		// Clear the cache after updating the index
-		// there could be new, updated or removed packages
-		i.cache.Purge()
+	if i.afterUpdateHook != nil {
+		i.afterUpdateHook(ctx)
 	}
 
 	metrics.StorageIndexerUpdateIndexSuccessTotal.Inc()