Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.
Question 1 · choose 1
A store's pages are cached by Cloud CDN behind a global external Application Load Balancer. Product 123 appears on its own page /p/123 and also on many category and search pages, such as /shoes/running and /search?q=trail. After a price change, every cached response that shows product 123 must be purged, and nothing else. The origin can add response headers. What should the team do?
ATag those responses with a Cache-Tag header such as product-123, then invalidate with the cacheTags matcher
BInvalidate the path /p/123* so that the product page and its variants are purged
CInvalidate the paths /p/123 and /search?q=* together in one request
DInvalidate the path /* so that every cached page is refilled from the origin with the current prices
Show the answer and why
ATag those responses with a Cache-Tag header such as product-123, then invalidate with the cacheTags matcher
Correct
The cacheTags matcher invalidates content grouped by the metadata in the Cache-Tag response header, whatever its URL.
BInvalidate the path /p/123* so that the product page and its variants are purged
Incorrect
A path prefix catches only URLs that start with /p/123; the category and search pages keep the old price.
CInvalidate the paths /p/123 and /search?q=* together in one request
Incorrect
A path matcher can't contain ? and still would miss the category pages.
DInvalidate the path /* so that every cached page is refilled from the origin with the current prices
Incorrect
Purging everything goes far beyond the product's pages and sends the whole site's cache fill to the backend.
Cloud CDN invalidation matchers include host, path, HTTP status, content type, backend and cache tags. When the content to purge is spread over unrelated URLs, tagging responses with a Cache-Tag header lets one request with the cacheTags matcher target exactly that group.
A team enables Cloud CDN on a backend service and leaves the cache mode at its default. Which cache mode is the backend service using?
AUSE_ORIGIN_HEADERS
BFORCE_CACHE_ALL
CCACHE_ALL_STATIC
DNo caching until a cache mode is chosen
Show the answer and why
AUSE_ORIGIN_HEADERS
Incorrect
This mode caches only responses with valid caching headers from the origin; it isn't the default.
BFORCE_CACHE_ALL
Incorrect
This mode caches all successful responses regardless of headers; it must be chosen explicitly.
CCACHE_ALL_STATIC
Correct
CACHE_ALL_STATIC is the default cache mode and automatically caches static content.
DNo caching until a cache mode is chosen
Incorrect
Cloud CDN applies a default cache mode as soon as it is enabled.
Cloud CDN has three cache modes: CACHE_ALL_STATIC (the default), which automatically caches static content; USE_ORIGIN_HEADERS, which follows the origin's caching headers; and FORCE_CACHE_ALL, which caches successful responses regardless of those headers.
Cloud CDN serves images from a private bucket in another provider's S3-compatible object storage, reached through an internet NEG. The bucket rejects anonymous requests. How should Cloud CDN be given access to this origin?
AConfigure private origin authentication with AWS Signature Version 4
BEnable signed URLs on the backend so that each client request carries a signature
CAdd the bucket to a Cloud Storage backend bucket with Standard Tier
DAllow-list Cloud CDN's client IP addresses in the bucket policy
Show the answer and why
AConfigure private origin authentication with AWS Signature Version 4
Correct
Private origin authentication signs requests to S3-compatible private origins with AWS Signature Version 4.
BEnable signed URLs on the backend so that each client request carries a signature
Incorrect
Signed URLs authorize clients to Cloud CDN; they don't authenticate Cloud CDN to the origin.
CAdd the bucket to a Cloud Storage backend bucket with Standard Tier
Incorrect
Backend buckets point to Cloud Storage buckets, not to another provider's storage.
DAllow-list Cloud CDN's client IP addresses in the bucket policy
Incorrect
Requests from allow-listed addresses still carry no credentials, which the bucket rejects; private origin authentication signs them instead.
For external origins that require authentication, such as private S3-compatible buckets behind an internet NEG, Cloud CDN supports private origin authentication: it signs its requests with AWS Signature Version 4 using credentials that you configure.
A release pipeline sends one Cloud CDN invalidation request per changed file and sometimes submits thousands of requests within a minute; many are rejected. What explains this?
AEach invalidation request can name only one object
BInvalidation requests are rate limited to 500 per minute per project
CInvalidations are accepted only while the backend is healthy
DEach URL map can hold only one pending invalidation at a time
Show the answer and why
AEach invalidation request can name only one object
Incorrect
Cloud CDN doesn't restrict the number of objects or their total size per request.
BInvalidation requests are rate limited to 500 per minute per project
Correct
The invalidation overview states this limit, and each request takes effect in about 10 seconds.
CInvalidations are accepted only while the backend is healthy
Incorrect
Backend health doesn't gate invalidation requests.
DEach URL map can hold only one pending invalidation at a time
Incorrect
The limit is a request rate per project, not one pending request.
Cloud CDN invalidation requests are rate limited: up to 500 per minute per project. Each takes effect in about 10 seconds, and a single request isn't limited in how many objects it covers, so path prefixes and other matchers can replace many single-file requests.