Extend Cloud Storage with Cloud Functions

You can trigger a function in response to the uploading, updating, or deleting of files and folders in Cloud Storage .

Examples in this page are based on a sample function that triggers when image files are uploaded to Cloud Storage . This sample function demonstrates how to access event attributes, how to download a file to a Cloud Functions instance, and other fundamentals of handling Cloud Storage events.

Import the required modules

To get started, import the module required for handling Cloud Storage events:

Node.js

   
 const 
  
 { 
 onObjectFinalized 
 } 
  
 = 
  
 require 
 ( 
 "firebase-functions/v2/storage" 
 ); 
   index 
 . 
 js

Python

  from 
  
 firebase_functions 
  
 import 
 storage_fn  main 
 
 
 . 
 py

To build out the full sample, also add the dependencies for the Firebase Admin SDK and image processing tools:

Node.js

   
 const 
  
 { 
 initializeApp 
 } 
  
 = 
  
 require 
 ( 
 "firebase-admin/app" 
 ); 
 const 
  
 { 
 getStorage 
 } 
  
 = 
  
 require 
 ( 
 "firebase-admin/storage" 
 ); 
 const 
  
 logger 
  
 = 
  
 require 
 ( 
 "firebase-functions/logger" 
 ); 
 const 
  
 path 
  
 = 
  
 require 
 ( 
 "path" 
 ); 
 // library for image resizing 
 const 
  
 sharp 
  
 = 
  
 require 
 ( 
 "sharp" 
 ); 
 initializeApp 
 (); 
   index 
 . 
 js

Python

  import 
  
 io 
 import 
  
 pathlib 
 from 
  
 PIL 
  
 import 
 Image 
 from 
  
 firebase_admin 
  
 import 
 initialize_app 
 initialize_app 
 () 
 from 
  
 firebase_admin 
  
 import 
 storage  main 
 
 
 . 
 py

Scope a Cloud Storage function

Use the following pattern to scope your function to a specific Cloud Storage bucket and set any desired options:

Node.js

  // scope handler to a specific bucket, using storage options parameter 
 exports 
 . 
 archivedopts 
  
 = 
  
 onObjectArchived 
 ({ 
  
 bucket 
 : 
  
 "myBucket" 
  
 }, 
  
 ( 
 event 
 ) 
  
 = 
>  
 { 
  
 //… 
 });

Python

  # Scope handler to a specific bucket using storage options parameter 
 @storage_fn 
 . 
 on_object_archived 
 ( 
 bucket 
 = 
 "myBucket" 
 ) 
 def 
  
 archived_bucket 
 ( 
 event 
 : 
 storage_fn 
 . 
 CloudEvent 
 [ 
 storage_fn 
 . 
 StorageObjectData 
 ]): 
 # ...

By contrast, the example thumbnail generator function is scoped to the default bucket for the project:

Node.js

 exports 
 . 
 generateThumbnail 
  
 = 
  
 onObjectFinalized 
 ({ 
 cpu 
 : 
  
 2 
 }, 
  
 async 
  
 ( 
 event 
 ) 
  
 = 
>  
 { 
 // ... 
 });

Python

  @storage_fn 
 . 
 on_object_archived 
 () 
 def 
  
 generatethumbnail 
 ( 
 event 
 : 
 storage_fn 
 . 
 CloudEvent 
 [ 
 storage_fn 
 . 
 StorageObjectData 
 ]): 
 # ...

Set the function location

A mismatch between locations can result in deployment failure. Also, distance between the location of a Cloud Storage bucket and the location of the function can create significant network latency. To avoid these situations, specify the function location so that it matches the bucket/trigger location in one of these ways:

Function location is the same as the trigger location
Function location is inside the trigger location (when the trigger region is dual/multi region)
Function may be in any location if the trigger region is set to us-central1

Handle Cloud Storage events

These handlers for responding to Cloud Storage events are available:

Node.js

onObjectArchived Only sent when a bucket has enabled object versioning . This event indicates that the live version of an object has become an archived version, either because it was archived or because it was overwritten by the upload of an object of the same name.
onObjectDeleted Sent when an object has been permanently deleted. This includes objects that are overwritten or are deleted as part of the bucket's lifecycle configuration . For buckets with object versioning enabled, this is not sent when an object is archived (see onArchive ), even if archival occurs via the storage.objects.delete method.
onObjectFinalized Sent when a new object (or a new generation of an existing object) is successfully created in the bucket. This includes copying or rewriting an existing object. A failed upload does not trigger this event.
onMetadataUpdated Sent when the metadata of an existing object changes.

Python

on_object_archived Only sent when a bucket has enabled object versioning . This event indicates that the live version of an object has become an archived version, either because it was archived or because it was overwritten by the upload of an object of the same name.
on_object_deleted Sent when an object has been permanently deleted. This includes objects that are overwritten or are deleted as part of the bucket's lifecycle configuration . For buckets with object versioning enabled, this is not sent when an object is archived (see onArchive ), even if archival occurs via the storage.objects.delete method.
on_object_finalized Sent when a new object (or a new generation of an existing object) is successfully created in the bucket. This includes copying or rewriting an existing object. A failed upload does not trigger this event.
on_metadata_updated Sent when the metadata of an existing object changes.

Access Cloud Storage object attributes

Cloud Functions exposes a number of Cloud Storage object attributes such as the object's size and content type for the file updated. The metageneration attribute is incremented whenever there's a change to the object's metadata. For new objects, the metageneration value is 1 .

Node.js

 const 
  
 fileBucket 
  
 = 
  
 event 
 . 
 data 
 . 
 bucket 
 ; 
  
 // Storage bucket containing the file. 
 const 
  
 filePath 
  
 = 
  
 event 
 . 
 data 
 . 
 name 
 ; 
  
 // File path in the bucket. 
 const 
  
 contentType 
  
 = 
  
 event 
 . 
 data 
 . 
 contentType 
 ; 
  
 // File content type.  index.js

Python

  bucket_name 
 = 
 event 
 . 
 data 
 . 
 bucket 
 file_path 
 = 
 pathlib 
 . 
 PurePath 
 ( 
 event 
 . 
 data 
 . 
 name 
 ) 
 content_type 
 = 
 event 
 . 
 data 
 . 
 content_type  main 
 
 
 . 
 py

The thumbnail generation sample uses some of these attributes to detect exit cases in which the function returns:

Node.js

 // Exit if this is triggered on a file that is not an image. 
 if 
  
 ( 
 ! 
 contentType 
 . 
 startsWith 
 ( 
 "image/" 
 )) 
  
 { 
  
 return 
  
 logger 
 . 
 log 
 ( 
 "This is not an image." 
 ); 
 } 
 // Exit if the image is already a thumbnail. 
 const 
  
 fileName 
  
 = 
  
 path 
 . 
 basename 
 ( 
 filePath 
 ); 
 if 
  
 ( 
 fileName 
 . 
 startsWith 
 ( 
 "thumb_" 
 )) 
  
 { 
  
 return 
  
 logger 
 . 
 log 
 ( 
 "Already a Thumbnail." 
 ); 
 } 
   index 
 . 
 js

Python

  # Exit if this is triggered on a file that is not an image. 
 if 
 not 
 content_type 
 or 
 not 
 content_type 
 . 
 startswith 
 ( 
 "image/" 
 ): 
 print 
 ( 
 f 
 "This is not an image. ( 
 { 
 content_type 
 } 
 )" 
 ) 
 return 
 # Exit if the image is already a thumbnail. 
 if 
 file_path 
 . 
 name 
 . 
 startswith 
 ( 
 "thumb_" 
 ): 
 print 
 ( 
 "Already a thumbnail." 
 ) 
 return  main 
 
 
 . 
 py

Download, transform, and upload a file

For some cases, it may not be necessary to download files from Cloud Storage . However, to perform intensive tasks such as generating a thumbnail image from a file stored in Cloud Storage , you need to download files to the functions instance—that is, the virtual machine that runs your code.

Using Cloud Functions together with image-processing programs like sharp for Node.js and Pillow for Python, you can perform manipulations on graphical image files. The following is an example of how to create a thumbnail image for an uploaded image file:

Node.js

  /** 
 * When an image is uploaded in the Storage bucket, 
 * generate a thumbnail automatically using sharp. 
 */ 
 exports 
 . 
 generateThumbnail 
  
 = 
  
 onObjectFinalized 
 ({ 
 cpu 
 : 
  
 2 
 }, 
  
 async 
  
 ( 
 event 
 ) 
  
 = 
>  
 { 
  
 const 
  
 fileBucket 
  
 = 
  
 event 
 . 
 data 
 . 
 bucket 
 ; 
  
 // Storage bucket containing the file. 
  
 const 
  
 filePath 
  
 = 
  
 event 
 . 
 data 
 . 
 name 
 ; 
  
 // File path in the bucket. 
  
 const 
  
 contentType 
  
 = 
  
 event 
 . 
 data 
 . 
 contentType 
 ; 
  
 // File content type. 
  
 // Exit if this is triggered on a file that is not an image. 
  
 if 
  
 ( 
 ! 
 contentType 
 . 
 startsWith 
 ( 
 "image/" 
 )) 
  
 { 
  
 return 
  
 logger 
 . 
 log 
 ( 
 "This is not an image." 
 ); 
  
 } 
  
 // Exit if the image is already a thumbnail. 
  
 const 
  
 fileName 
  
 = 
  
 path 
 . 
 basename 
 ( 
 filePath 
 ); 
  
 if 
  
 ( 
 fileName 
 . 
 startsWith 
 ( 
 "thumb_" 
 )) 
  
 { 
  
 return 
  
 logger 
 . 
 log 
 ( 
 "Already a Thumbnail." 
 ); 
  
 } 
  
 // Download file into memory from bucket. 
  
 const 
  
 bucket 
  
 = 
  
 getStorage 
 (). 
 bucket 
 ( 
 fileBucket 
 ); 
  
 const 
  
 downloadResponse 
  
 = 
  
 await 
  
 bucket 
 . 
 file 
 ( 
 filePath 
 ). 
 download 
 (); 
  
 const 
  
 imageBuffer 
  
 = 
  
 downloadResponse 
 [ 
 0 
 ]; 
  
 logger 
 . 
 log 
 ( 
 "Image downloaded!" 
 ); 
  
 // Generate a thumbnail using sharp. 
  
 const 
  
 thumbnailBuffer 
  
 = 
  
 await 
  
 sharp 
 ( 
 imageBuffer 
 ). 
 resize 
 ({ 
  
 width 
 : 
  
 200 
 , 
  
 height 
 : 
  
 200 
 , 
  
 withoutEnlargement 
 : 
  
 true 
 , 
  
 }). 
 toBuffer 
 (); 
  
 logger 
 . 
 log 
 ( 
 "Thumbnail created" 
 ); 
  
 // Prefix 'thumb_' to file name. 
  
 const 
  
 thumbFileName 
  
 = 
  
 `thumb_ 
 ${ 
 fileName 
 } 
 ` 
 ; 
  
 const 
  
 thumbFilePath 
  
 = 
  
 path 
 . 
 join 
 ( 
 path 
 . 
 dirname 
 ( 
 filePath 
 ), 
  
 thumbFileName 
 ); 
  
 // Upload the thumbnail. 
  
 const 
  
 metadata 
  
 = 
  
 { 
 contentType 
 : 
  
 contentType 
 }; 
  
 await 
  
 bucket 
 . 
 file 
 ( 
 thumbFilePath 
 ). 
 save 
 ( 
 thumbnailBuffer 
 , 
  
 { 
  
 metadata 
 : 
  
 metadata 
 , 
  
 }); 
  
 return 
  
 logger 
 . 
 log 
 ( 
 "Thumbnail uploaded!" 
 ); 
 }); 
   index 
 . 
 js

Download the file to a temporary directory on your Cloud Functions instance. In this location, you can process the file as needed and then upload to Cloud Storage . When performing asynchronous tasks, make sure you return a JavaScript promise in your callback.

Python

  @storage_fn 
 . 
 on_object_finalized 
 () 
 def 
  
 generatethumbnail 
 ( 
 event 
 : 
 storage_fn 
 . 
 CloudEvent 
 [ 
 storage_fn 
 . 
 StorageObjectData 
 ]): 
  
 """When an image is uploaded in the Storage bucket, generate a thumbnail 
 automatically using Pillow.""" 
 bucket_name 
 = 
 event 
 . 
 data 
 . 
 bucket 
 file_path 
 = 
 pathlib 
 . 
 PurePath 
 ( 
 event 
 . 
 data 
 . 
 name 
 ) 
 content_type 
 = 
 event 
 . 
 data 
 . 
 content_type 
 # Exit if this is triggered on a file that is not an image. 
 if 
 not 
 content_type 
 or 
 not 
 content_type 
 . 
 startswith 
 ( 
 "image/" 
 ): 
 print 
 ( 
 f 
 "This is not an image. ( 
 { 
 content_type 
 } 
 )" 
 ) 
 return 
 # Exit if the image is already a thumbnail. 
 if 
 file_path 
 . 
 name 
 . 
 startswith 
 ( 
 "thumb_" 
 ): 
 print 
 ( 
 "Already a thumbnail." 
 ) 
 return 
 bucket 
 = 
 storage 
 . 
 bucket 
 ( 
 bucket_name 
 ) 
 image_blob 
 = 
 bucket 
 . 
 blob 
 ( 
 str 
 ( 
 file_path 
 )) 
 image_bytes 
 = 
 image_blob 
 . 
 download_as_bytes 
 () 
 image 
 = 
 Image 
 . 
 open 
 ( 
 io 
 . 
 BytesIO 
 ( 
 image_bytes 
 )) 
 image 
 . 
 thumbnail 
 (( 
 200 
 , 
 200 
 )) 
 thumbnail_io 
 = 
 io 
 . 
 BytesIO 
 () 
 image 
 . 
 save 
 ( 
 thumbnail_io 
 , 
 format 
 = 
 "png" 
 ) 
 thumbnail_path 
 = 
 file_path 
 . 
 parent 
 / 
 pathlib 
 . 
 PurePath 
 ( 
 f 
 "thumb_ 
 { 
 file_path 
 . 
 stem 
 } 
 .png" 
 ) 
 thumbnail_blob 
 = 
 bucket 
 . 
 blob 
 ( 
 str 
 ( 
 thumbnail_path 
 )) 
 thumbnail_blob 
 . 
 upload_from_string 
 ( 
 thumbnail_io 
 . 
 getvalue 
 (), 
 content_type 
 = 
 "image/png" 
 ) 
   main 
 . 
 py

This code creates a 200x200 thumbnail for the image saved in a temporary directory, then uploads it back to Cloud Storage .

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License , and code samples are licensed under the Apache 2.0 License . For details, see the Google Developers Site Policies . Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025-09-05 UTC.

Create a Mobile Website

View Site in Mobile | Classic

Share by:

Extend Cloud Storage with Cloud Functions Stay organized with collections Save and categorize content based on your preferences.

Import the required modules

Node.js

Python

Node.js

Python

Scope a Cloud Storage function

Node.js

Python

Node.js

Python

Set the function location

Handle Cloud Storage events

Node.js

Python

Access Cloud Storage object attributes

Node.js

Python

Node.js

Python

Download, transform, and upload a file

Node.js

Python

Extend Cloud Storage with Cloud Functions