This page discusses the commonly-used metadata fields that are stored along with objects in Cloud Storage.
Introduction
Objects stored in Cloud Storage have metadata associated with them.
Metadata identifies properties of the object, as well as specifies how the
object should be handled when it's accessed. Metadata exists as key:value
pairs. For example, the storage class of an object is represented
by the metadata entry storageClass:STANDARD
. storageClass
is the key
for the metadata, and all objects have such a key associated with them.
STANDARD
specifies the value this specific object has, and the value varies
from object to object.
The mutability of metadata varies: some metadata you can edit at any time, some
metadata you can only set at the time the object is created, and some metadata
you can only view. For example, you can edit the value of the Cache-Control
metadata at any time, but you can only assign the storageClass
metadata when
the object is created or rewritten, and you cannot directly edit the value
for the generation
metadata, though the generation
value changes when
the object is replaced.
Editable metadata
There are two categories of metadata that users can change for objects:
Fixed-key metadata: Metadata whose keys are set, but for which you can specify a value.
Custom metadata: Metadata that you add by specifying both a key and a value associated with the key.
When editing metadata, you should generally avoid non-ascii characters, because they are not permitted in HTTP headers, which the XML API uses.
Fixed-key metadata
You can edit the following metadata for objects, though you must have sufficient permission to do so:
- Access control metadata
- Cache-Control
- Content-Disposition
- Content-Encoding
- Content-Language
- Content-Type
- Custom-Time
- Object holds
- Retention configuration
Access control metadata
Cloud Storage uses Identity and Access Management (IAM) and Access Control Lists (ACLs) to control access to objects. Use these links to learn about these access control methods and associated metadata.
Cache-Control
The Cache-Control
metadata can specify two different aspects
of how data is served from Cloud Storage: whether the data can be
cached and whether the data can be transformed.
Caching data
The Cache-Control
metadata lets you control whether and for how long caches
are allowed to cache your objects, which can then be served to satisfy future
requests. Caches can include browser and Internet caches, as well as
Cloud Storage built-in caching.
If an applicable object doesn't have a Cache-Control
metadata entry,
Cloud Storage uses the default value of:
public, max-age=3600
, if the object is not encrypted using a customer-managed encryption key or stored within a Virtual Private Cloud service perimeter.no-cache, no-store, max-age=0
, if the object is encrypted using a customer-managed encryption key.private, max-age=0
, if the object is stored within a Virtual Private Cloud service perimeter.no-cache, no-store, max-age=0, must-revalidate
, if the object is stored within a Virtual Private Cloud service perimeter and encrypted using a customer-managed encryption key.
If you allow caching, downloads might continue to receive earlier versions of
an object, even after uploading a newer version. This is because such earlier
versions remain "fresh" in caches for a period of time determined by
max-age
. Additionally, because objects can be cached at various places on the
Internet, there is no way to force a cached object to expire globally. This
means that if you revoke public access to an object, the object can still be
served from a cache, depending on when it was last accessed and its
Cache-Control
setting. For example, if your object was served with a
Cache-Control
of public, max-age=3600
, it can persist in a cache for an
hour. If you want to prevent serving cached versions of publicly readable
objects, set Cache-Control: no-store
on the object.
If you need greater control over cache behavior, you can configure Cloud CDN in front of your bucket.
Transforming data
Cache-Control
metadata also lets you serve objects as they are stored,
without applying any transformations to the data, such as removing gzip
content-encoding for incompatible clients. To serve an object as-is, set
Cache-Control:no-transform
.
Content-Disposition
The Content-Disposition
metadata specifies presentation
information about the data being transmitted. Setting Content-Disposition
allows you to control presentation style of the content, for example
determining whether an attachment should be automatically displayed or whether
some form of action from the user should be required to open it. See
https://datatracker.ietf.org/doc/html/rfc6266 for the
Content-Disposition
specification.
Content-Encoding
The Content-Encoding
metadata can be used to indicate that an object is
compressed, while still maintaining the object's underlying Content-Type
.
For example, a text file that is gzip compressed can have the fact that
it's a text file indicated in Content-Type
and the fact that it's
gzip compressed indicated in Content-Encoding
. You should ensure that
files are, in fact, compressed using the specified Content-Encoding
before
uploading them, or else unexpected behavior can occur when attempting to
download the objects. For more information, see the Transcoding page.
For compressible content, such as text, using Content-Encoding: gzip
saves
network and storage costs and improves content serving performance. However,
for content that is already inherently compressed, such as archives and many
media formats, applying another level of compression and marking it in the
Content-Encoding
metadata is typically detrimental to both object size and
performance and should be avoided.
Content-Language
The Content-Language
metadata indicates the language(s) that
the object is intended for. Refer to ISO 639-1 language codes
for typical values of this metadata.
Cloud Storage supports Content-Language
values up to 100 characters
in length.
Content-Type
The most commonly set metadata is Content-Type
(also known as media type),
which lets browsers render the object properly. All objects have a value
specified in their Content-Type
metadata, but this value does not have to
match the underlying type of the object. For example, if the Content-Type
is
not specified by the uploader and cannot be determined, it is set to
application/octet-stream
or application/x-www-form-urlencoded
, depending on
how you uploaded the object. For a list of valid content types, see the
IANA Media Types page.
Custom-Time
The Custom-Time
metadata is a user-specified date and time represented in
the RFC 3339 format YYYY-MM-DD'T'HH:MM:SS.SS'Z'
or
YYYY-MM-DD'T'HH:MM:SS'Z'
when milliseconds are zero. This metadata is
typically set in order to use the DaysSinceCustomTime
condition in
Object Lifecycle Management.
You cannot remove Custom-Time
once it's been set on an object. Additionally,
the value for Custom-Time
cannot decrease. That is, you cannot set
Custom-Time
to be an earlier date/time than the existing Custom-Time
. You
can, however, effectively remove or reset the Custom-Time
by
rewriting the object.
Object holds
Use metadata flags to place object holds, which prevent objects from being deleted or replaced. For more information, see the Object holds page.
Retention configuration
When present, an object's retention configuration defines a date and time prior to which the object cannot be deleted or replaced. For more information, see Object Retention Lock.
Custom metadata
Custom metadata is metadata for which you define both the key and the value. To
create custom metadata, you specify both a key and a value. After you have
created a custom metadata key:value
pair, you can delete the key or change the
value.
Custom metadata is subject to a size limit and incurs storage costs.
The Viewing and Editing Metadata page includes information about setting custom metadata.
The x-goog-meta-
prefix
The XML API sets and retrieves object metadata using request headers, and
the JSON API allows setting custom metadata in the final request of a
resumable upload by using request headers. In order to clearly distinguish
custom metadata headers from standard request headers, both APIs prefix such
custom metadata headers with x-goog-meta-
.
Non-editable metadata
Some metadata cannot be edited directly. This metadata is set at the time of object creation or rewrite. As part of the object creation or rewrite, you can set some such metadata, such as the storage class of the object or customer-managed encryption keys. Other metadata is automatically added and can only be viewed, such the generation number of the object or the time of creation.
Generation and metageneration numbers
As part of its metadata, every Cloud Storage object has a numeric
generation
property and a numeric metageneration
property that uniquely
identifies the object:
Property | Description |
---|---|
generation |
Identifies the version of an object, and exists for every object, regardless of whether a bucket uses Object Versioning.
|
metageneration |
Identifies the metadata version, and increases every time the metadata of a given generation is updated.
|
The generation
and metageneration
properties are used in the following
circumstances:
When using preconditions in requests: Preconditions cause the request to fail if the precondition isn't satisfied. Failing in this manner prevents the request from applying to an unexpected version of an object, such as retrieving the wrong object data or modifying the wrong state of an object's metadata.
When listing, accessing, restoring, and deleting noncurrent object versions: Noncurrent object versions are specifically relevant in buckets that use or previously used Object Versioning.
Checksums
Checksums are metadata calculated from the associated object's data. Checksums are used for validating that object data isn't corrupted. Cloud Storage objects have several checksum metadata fields.
CRC32C
All Cloud Storage objects have a CRC32C hash. Libraries for computing CRC32C include:
- Google's CRC32C for C++
- hash/crc32 for Go
- GoogleAPIs Guava for Java
- google-crc32c for Python
- digest-crc in Ruby
The Base64 encoded CRC32C is in big-endian byte order.
MD5
Cloud Storage objects have an MD5 hash if they meet the following criteria:
- The object is not a composite object
- The object was not uploaded using an XML API multipart upload
This hash only applies to a complete object, so it cannot be used to integrity check partial downloads caused by performing a range GET.
ETags
All Cloud Storage objects have an ETag. However, the same object can have a different ETag value when it's requested from the XML API compared to the JSON API. In most cases, users should make no assumptions about the value used in an ETag except that it changes whenever the underlying data or metadata changes, per the specification.
An object's ETag header returns the object's MD5 value if all the following are true:
- The request is being made through the XML API
- The object only uses Google-managed encryption keys for server-side encryption
- The object is not a composite object and was not uploaded using an XML API multipart upload
Modification time
As part of its metadata, every Cloud Storage object has an updated
property that indicates the last time the object's metadata was modified. The
updated
time is set initially to the object's creation time and then changes
whenever any metadata of the object changes. This includes changes made by a
requester, such as modifying custom metadata, as well as changes made by
Cloud Storage on behalf of a requester, such as changing the storage
class based on an Object Lifecycle Configuration.
What's next
- View and edit object metadata.
- Learn about available storage classes.
- For a detailed description of all object metadata fields available in the JSON API, see the Objects reference documentation for JSON.
- Learn about Storage Insights inventory reports, which allow you to get the metadata for all objects in a bucket at once.