[release-16.0] Address vttablet memory usage with backups to Azure Bl…

…ob Service (#13770) (#13774)

Signed-off-by: Shlomi Noach <2607934+shlomi-noach@users.noreply.github.com>
Co-authored-by: vitess-bot[bot] <108069721+vitess-bot[bot]@users.noreply.github.com>
Co-authored-by: Shlomi Noach <2607934+shlomi-noach@users.noreply.github.com>

go/flags/endtoend/vtbackup.txt

@@ -3,6 +3,7 @@ Usage of vtbackup:
       --alsologtostderr                                 log to standard error as well as files
       --azblob_backup_account_key_file string           Path to a file containing the Azure Storage account key; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_KEY will be used as the key itself (NOT a file path).
       --azblob_backup_account_name string               Azure Storage Account name for backups; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_NAME will be used.
+      --azblob_backup_buffer_size int                   The memory buffer size to use in bytes, per file or stripe, when streaming to Azure Blob Service. (default 104857600)
       --azblob_backup_container_name string             Azure Blob Container Name.
       --azblob_backup_parallelism int                   Azure Blob operation parallelism (requires extra memory when increased). (default 1)
       --azblob_backup_storage_root string               Root prefix for all backup-related Azure Blobs; this should exclude both initial and trailing '/' (e.g. just 'a/b' not '/a/b/').

go/flags/endtoend/vtctld.txt

@@ -3,6 +3,7 @@ Usage of vtctld:
       --alsologtostderr                                                  log to standard error as well as files
       --azblob_backup_account_key_file string                            Path to a file containing the Azure Storage account key; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_KEY will be used as the key itself (NOT a file path).
       --azblob_backup_account_name string                                Azure Storage Account name for backups; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_NAME will be used.
+      --azblob_backup_buffer_size int                                    The memory buffer size to use in bytes, per file or stripe, when streaming to Azure Blob Service. (default 104857600)
       --azblob_backup_container_name string                              Azure Blob Container Name.
       --azblob_backup_parallelism int                                    Azure Blob operation parallelism (requires extra memory when increased). (default 1)
       --azblob_backup_storage_root string                                Root prefix for all backup-related Azure Blobs; this should exclude both initial and trailing '/' (e.g. just 'a/b' not '/a/b/').

go/flags/endtoend/vttablet.txt

@@ -4,6 +4,7 @@ Usage of vttablet:
       --app_pool_size int                                                Size of the connection pool for app connections (default 40)
       --azblob_backup_account_key_file string                            Path to a file containing the Azure Storage account key; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_KEY will be used as the key itself (NOT a file path).
       --azblob_backup_account_name string                                Azure Storage Account name for backups; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_NAME will be used.
+      --azblob_backup_buffer_size int                                    The memory buffer size to use in bytes, per file or stripe, when streaming to Azure Blob Service. (default 104857600)
       --azblob_backup_container_name string                              Azure Blob Container Name.
       --azblob_backup_parallelism int                                    Azure Blob operation parallelism (requires extra memory when increased). (default 1)
       --azblob_backup_storage_root string                                Root prefix for all backup-related Azure Blobs; this should exclude both initial and trailing '/' (e.g. just 'a/b' not '/a/b/').

go/vt/mysqlctl/azblobbackupstorage/azblob.go

@@ -52,13 +52,15 @@ var (
 	storageRoot string
 
 	azBlobParallelism int
+	azBlobBufferSize  = 100 << (10 * 2) // 100 MiB
 )
 
 func registerFlags(fs *pflag.FlagSet) {
 	fs.StringVar(&accountName, "azblob_backup_account_name", "", "Azure Storage Account name for backups; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_NAME will be used.")
 	fs.StringVar(&accountKeyFile, "azblob_backup_account_key_file", "", "Path to a file containing the Azure Storage account key; if this flag is unset, the environment variable VT_AZBLOB_ACCOUNT_KEY will be used as the key itself (NOT a file path).")
 	fs.StringVar(&containerName, "azblob_backup_container_name", "", "Azure Blob Container Name.")
 	fs.StringVar(&storageRoot, "azblob_backup_storage_root", "", "Root prefix for all backup-related Azure Blobs; this should exclude both initial and trailing '/' (e.g. just 'a/b' not '/a/b/').")
+	fs.IntVar(&azBlobBufferSize, "azblob_backup_buffer_size", azBlobBufferSize, "The memory buffer size to use in bytes, per file or stripe, when streaming to Azure Blob Service.")
 	fs.IntVar(&azBlobParallelism, "azblob_backup_parallelism", 1, "Azure Blob operation parallelism (requires extra memory when increased).")
 }
 
@@ -218,7 +220,7 @@ func (bh *AZBlobBackupHandle) AddFile(ctx context.Context, filename string, file
 	go func() {
 		defer bh.waitGroup.Done()
 		_, err := azblob.UploadStreamToBlockBlob(bh.ctx, reader, blockBlobURL, azblob.UploadStreamToBlockBlobOptions{
-			BufferSize: azblob.BlockBlobMaxStageBlockBytes,
+			BufferSize: azBlobBufferSize,
 			MaxBuffers: azBlobParallelism,
 		})
 		if err != nil {