Kafka notify: support batched commits for queue store (#20377)

The items will be saved per target batch and will
be committed to the queue store when the batch is full

Also, periodically commit the batched items to the queue store
based on configured commit_timeout; default is 30s;

Bonus: compress queue store multi writes
This commit is contained in:
Praveen raj Mani
2024-09-07 04:36:30 +05:30
committed by GitHub
parent 0f1e8db4c5
commit 261111e728
20 changed files with 907 additions and 397 deletions
+86 -63
View File
@@ -18,100 +18,123 @@
package store
import (
"context"
"errors"
"fmt"
"sync"
"time"
)
// ErrBatchFull indicates that the batch is full
var ErrBatchFull = errors.New("batch is full")
type key interface {
string | int | int64
}
const defaultCommitTimeout = 30 * time.Second
// Batch represents an ordered batch
type Batch[K key, T any] struct {
keys []K
items map[K]T
limit uint32
type Batch[I any] struct {
items []I
limit uint32
store Store[I]
quitCh chan struct{}
sync.Mutex
}
// BatchConfig represents the batch config
type BatchConfig[I any] struct {
Limit uint32
Store Store[I]
CommitTimeout time.Duration
Log logger
}
// Add adds the item to the batch
func (b *Batch[K, T]) Add(key K, item T) error {
func (b *Batch[I]) Add(item I) error {
b.Lock()
defer b.Unlock()
if b.isFull() {
return ErrBatchFull
if b.store == nil {
return ErrBatchFull
}
// commit batch to store
if err := b.commit(); err != nil {
return err
}
}
if _, ok := b.items[key]; !ok {
b.keys = append(b.keys, key)
}
b.items[key] = item
b.items = append(b.items, item)
return nil
}
// GetAll fetches the items and resets the batch
// Returned items are not referenced by the batch
func (b *Batch[K, T]) GetAll() (orderedKeys []K, orderedItems []T, err error) {
b.Lock()
defer b.Unlock()
orderedKeys = append([]K(nil), b.keys...)
for _, key := range orderedKeys {
item, ok := b.items[key]
if !ok {
err = fmt.Errorf("item not found for the key: %v; should not happen;", key)
return
}
orderedItems = append(orderedItems, item)
delete(b.items, key)
}
b.keys = b.keys[:0]
return
}
// GetByKey will get the batch item by the provided key
func (b *Batch[K, T]) GetByKey(key K) (T, bool) {
b.Lock()
defer b.Unlock()
item, ok := b.items[key]
return item, ok
}
// Len returns the no of items in the batch
func (b *Batch[K, T]) Len() int {
func (b *Batch[_]) Len() int {
b.Lock()
defer b.Unlock()
return len(b.keys)
return len(b.items)
}
// IsFull checks if the batch is full or not
func (b *Batch[K, T]) IsFull() bool {
b.Lock()
defer b.Unlock()
return b.isFull()
}
func (b *Batch[K, T]) isFull() bool {
func (b *Batch[_]) isFull() bool {
return len(b.items) >= int(b.limit)
}
// NewBatch creates a new batch
func NewBatch[K key, T any](limit uint32) *Batch[K, T] {
return &Batch[K, T]{
keys: make([]K, 0, limit),
items: make(map[K]T, limit),
limit: limit,
func (b *Batch[I]) commit() error {
switch len(b.items) {
case 0:
return nil
case 1:
_, err := b.store.Put(b.items[0])
return err
default:
}
if _, err := b.store.PutMultiple(b.items); err != nil {
return err
}
b.items = make([]I, 0, b.limit)
return nil
}
// Close commits the pending items and quits the goroutines
func (b *Batch[I]) Close() error {
defer func() {
close(b.quitCh)
}()
b.Lock()
defer b.Unlock()
return b.commit()
}
// NewBatch creates a new batch
func NewBatch[I any](config BatchConfig[I]) *Batch[I] {
if config.CommitTimeout == 0 {
config.CommitTimeout = defaultCommitTimeout
}
quitCh := make(chan struct{})
batch := &Batch[I]{
items: make([]I, 0, config.Limit),
limit: config.Limit,
store: config.Store,
quitCh: quitCh,
}
if batch.store != nil {
go func() {
commitTicker := time.NewTicker(config.CommitTimeout)
defer commitTicker.Stop()
for {
select {
case <-commitTicker.C:
case <-batch.quitCh:
return
}
batch.Lock()
err := batch.commit()
batch.Unlock()
if err != nil {
config.Log(context.Background(), err, "")
}
}
}()
}
return batch
}