CBG-4032: memory based rev cache implementation (#7040)

* CBG-4032: memory based rev cache implementation

* make test pass with default collection + missing return param

* more issues with default collection test

* touch ups on comment + remove unused function

* updated for review

* further tidy up

* lint fix after refector test

* fixes from rebase

* updates based off review + some more refactoring

base/stats.go

@@ -434,6 +434,8 @@ type CacheStats struct {
 	RevisionCacheHits *SgwIntStat `json:"rev_cache_hits"`
 	// The total number of revision cache misses.
 	RevisionCacheMisses *SgwIntStat `json:"rev_cache_misses"`
+	// Total memory used by the rev cache
+	RevisionCacheTotalMemory *SgwIntStat `json:"revision_cache_total_memory"`
 	// The current length of the pending skipped sequence slice.
 	SkippedSeqLen *SgwIntStat `json:"skipped_seq_len"`
 	// The current capacity of the skipped sequence slice
@@ -1339,6 +1341,10 @@ func (d *DbStats) initCacheStats() error {
 	if err != nil {
 		return err
 	}
+	resUtil.RevisionCacheTotalMemory, err = NewIntStat(SubsystemCacheKey, "revision_cache_total_memory", StatUnitNoUnits, RevCacheMemoryDesc, StatAddedVersion3dot2dot1, StatDeprecatedVersionNotDeprecated, StatStabilityCommitted, labelKeys, labelVals, prometheus.GaugeValue, 0)
+	if err != nil {
+		return err
+	}
 	resUtil.SkippedSeqLen, err = NewIntStat(SubsystemCacheKey, "skipped_seq_len", StatUnitNoUnits, SkippedSeqLengthDesc, StatAddedVersion3dot0dot0, StatDeprecatedVersionNotDeprecated, StatStabilityCommitted, labelKeys, labelVals, prometheus.GaugeValue, 0)
 	if err != nil {
 		return err
@@ -1388,6 +1394,7 @@ func (d *DbStats) unregisterCacheStats() {
 	prometheus.Unregister(d.CacheStats.RevisionCacheBypass)
 	prometheus.Unregister(d.CacheStats.RevisionCacheHits)
 	prometheus.Unregister(d.CacheStats.RevisionCacheMisses)
+	prometheus.Unregister(d.CacheStats.RevisionCacheTotalMemory)
 	prometheus.Unregister(d.CacheStats.SkippedSeqLen)
 	prometheus.Unregister(d.CacheStats.ViewQueries)
 }

base/stats_descriptions.go

@@ -129,6 +129,8 @@ const (
 	RevCacheMissesDesc = "The total number of revision cache misses. This metric can be used to calculate the ratio of revision cache misses: " +
 		"Rev Cache Miss Ratio = rev_cache_misses / (rev_cache_hits + rev_cache_misses)"
 
+	RevCacheMemoryDesc = "The approximation of total memory taken up by rev cache for documents. This is measured by the raw document body, the channels allocated to a document and its revision history."
+
 	SkippedSeqLengthDesc = "The current length of the pending skipped sequence slice."
 
 	SkippedSeqCapDesc = "The current capacity of the skipped sequence slice."

db/database_test.go

@@ -402,8 +402,13 @@ func TestGetRemovedAsUser(t *testing.T) {
 	// Manually remove the temporary backup doc from the bucket
 	// Manually flush the rev cache
 	// After expiry from the rev cache and removal of doc backup, try again
-	cacheHitCounter, cacheMissCounter, cacheNumItems := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems
-	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(DefaultRevisionCacheShardCount, DefaultRevisionCacheSize, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems)
+	cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStat := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems, db.DatabaseContext.DbStats.Cache().RevisionCacheTotalMemory
+	cacheOptions := &RevisionCacheOptions{
+		MaxBytes:     0,
+		MaxItemCount: DefaultRevisionCacheSize,
+		ShardCount:   DefaultRevisionCacheShardCount,
+	}
+	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(cacheOptions, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStat)
 	err = collection.PurgeOldRevisionJSON(ctx, "doc1", rev2id)
 	assert.NoError(t, err, "Purge old revision JSON")
 
@@ -754,8 +759,13 @@ func TestGetRemoved(t *testing.T) {
 	// Manually remove the temporary backup doc from the bucket
 	// Manually flush the rev cache
 	// After expiry from the rev cache and removal of doc backup, try again
-	cacheHitCounter, cacheMissCounter, cacheNumItems := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems
-	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(DefaultRevisionCacheShardCount, DefaultRevisionCacheSize, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems)
+	cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStat := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems, db.DatabaseContext.DbStats.Cache().RevisionCacheTotalMemory
+	cacheOptions := &RevisionCacheOptions{
+		MaxBytes:     0,
+		MaxItemCount: DefaultRevisionCacheSize,
+		ShardCount:   DefaultRevisionCacheShardCount,
+	}
+	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(cacheOptions, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStat)
 	err = collection.PurgeOldRevisionJSON(ctx, "doc1", rev2id)
 	assert.NoError(t, err, "Purge old revision JSON")
 
@@ -823,8 +833,13 @@ func TestGetRemovedAndDeleted(t *testing.T) {
 	// Manually remove the temporary backup doc from the bucket
 	// Manually flush the rev cache
 	// After expiry from the rev cache and removal of doc backup, try again
-	cacheHitCounter, cacheMissCounter, cacheNumItems := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems
-	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(DefaultRevisionCacheShardCount, DefaultRevisionCacheSize, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems)
+	cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStats := db.DatabaseContext.DbStats.Cache().RevisionCacheHits, db.DatabaseContext.DbStats.Cache().RevisionCacheMisses, db.DatabaseContext.DbStats.Cache().RevisionCacheNumItems, db.DatabaseContext.DbStats.Cache().RevisionCacheTotalMemory
+	cacheOptions := &RevisionCacheOptions{
+		MaxBytes:     0,
+		MaxItemCount: DefaultRevisionCacheSize,
+		ShardCount:   DefaultRevisionCacheShardCount,
+	}
+	collection.dbCtx.revisionCache = NewShardedLRURevisionCache(cacheOptions, backingStoreMap, cacheHitCounter, cacheMissCounter, cacheNumItems, memoryCacheStats)
 	err = collection.PurgeOldRevisionJSON(ctx, "doc1", rev2id)
 	assert.NoError(t, err, "Purge old revision JSON")
 

db/revision_cache_interface.go

@@ -68,31 +68,33 @@ func NewRevisionCache(cacheOptions *RevisionCacheOptions, backingStores map[uint
 		cacheOptions = DefaultRevisionCacheOptions()
 	}
 
-	if cacheOptions.Size == 0 {
+	if cacheOptions.MaxItemCount == 0 {
 		bypassStat := cacheStats.RevisionCacheBypass
 		return NewBypassRevisionCache(backingStores, bypassStat)
 	}
 
 	cacheHitStat := cacheStats.RevisionCacheHits
 	cacheMissStat := cacheStats.RevisionCacheMisses
 	cacheNumItemsStat := cacheStats.RevisionCacheNumItems
+	cacheMemoryStat := cacheStats.RevisionCacheTotalMemory
 
 	if cacheOptions.ShardCount > 1 {
-		return NewShardedLRURevisionCache(cacheOptions.ShardCount, cacheOptions.Size, backingStores, cacheHitStat, cacheMissStat, cacheNumItemsStat)
+		return NewShardedLRURevisionCache(cacheOptions, backingStores, cacheHitStat, cacheMissStat, cacheNumItemsStat, cacheMemoryStat)
 	}
 
-	return NewLRURevisionCache(cacheOptions.Size, backingStores, cacheHitStat, cacheMissStat, cacheNumItemsStat)
+	return NewLRURevisionCache(cacheOptions, backingStores, cacheHitStat, cacheMissStat, cacheNumItemsStat, cacheMemoryStat)
 }
 
 type RevisionCacheOptions struct {
-	Size       uint32
-	ShardCount uint16
+	MaxItemCount uint32
+	MaxBytes     int64
+	ShardCount   uint16
 }
 
 func DefaultRevisionCacheOptions() *RevisionCacheOptions {
 	return &RevisionCacheOptions{
-		Size:       DefaultRevisionCacheSize,
-		ShardCount: DefaultRevisionCacheShardCount,
+		MaxItemCount: DefaultRevisionCacheSize,
+		ShardCount:   DefaultRevisionCacheShardCount,
 	}
 }
 
@@ -163,7 +165,8 @@ type DocumentRevision struct {
 	Attachments AttachmentsMeta
 	Delta       *RevisionDelta
 	Deleted     bool
-	Removed     bool // True if the revision is a removal.
+	Removed     bool  // True if the revision is a removal.
+	MemoryBytes int64 // storage of the doc rev bytes measurement, includes size of delta when present too
 }
 
 // MutableBody returns a deep copy of the given document revision as a plain body (without any special properties)
@@ -317,17 +320,20 @@ type RevisionDelta struct {
 	ToChannels            base.Set                // Full list of channels for the to revision
 	RevisionHistory       []string                // Revision history from parent of ToRevID to source revID, in descending order
 	ToDeleted             bool                    // Flag if ToRevID is a tombstone
+	totalDeltaBytes       int64                   // totalDeltaBytes is the total bytes for channels, revisions and body on the delta itself
 }
 
 func newRevCacheDelta(deltaBytes []byte, fromRevID string, toRevision DocumentRevision, deleted bool, toRevAttStorageMeta []AttachmentStorageMeta) RevisionDelta {
-	return RevisionDelta{
+	revDelta := RevisionDelta{
 		ToRevID:               toRevision.RevID,
 		DeltaBytes:            deltaBytes,
 		AttachmentStorageMeta: toRevAttStorageMeta,
 		ToChannels:            toRevision.Channels,
 		RevisionHistory:       toRevision.History.parseAncestorRevisions(fromRevID),
 		ToDeleted:             deleted,
 	}
+	revDelta.CalculateDeltaBytes()
+	return revDelta
 }
 
 // This is the RevisionCacheLoaderFunc callback for the context's RevisionCache.