Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team deploys a product catalog microservice on Cloud Run and exposes it to external partners through an Apigee API proxy. You must configure traffic management policies in the proxy to meet these requirements:
GET responses across all authorized partners, while excluding write operations and error responses.Which Apigee policy configuration should you implement?
This approach relies on an asynchronous Quota policy without SpikeArrest, omits cache population safeguards, and keys cached entries by client_id.
Quota with asynchronous synchronization reduces overhead, but it allows client apps to exceed their configured limit slightly before counters sync across nodes. Only a synchronous Quota policy guarantees accurate counting across all regions.SpikeArrest leaves the Cloud Run service exposed to sudden micro-bursts from misbehaving clients, which can overload backend instances or exhaust partner quotas rapidly.<SkipCachePopulation> allows unsuccessful responses (such as 4xx or 5xx errors) and non-GET write requests to enter the cache, causing dirty reads and broken state.client_id in the cache key stores separate copies of the identical product catalog for every partner.Quota: Trades strict accuracy for lower synchronization overhead in very high-throughput scenarios.ResponseCache: Requires explicit <SkipCachePopulation> conditions so only successful 200 GET responses are persisted.You did not need burst protection, slight quota overages were acceptable under your partner SLAs, and each partner received distinct, private responses requiring client_id isolation.
This configuration swaps the intended roles of the Quota and SpikeArrest policies, places ResponseCache in the wrong flow location, and unnecessarily fragments the cache using client_id.
Quota is designed for long-term contractual enforcement rather than sudden burst mitigation. Placing Quota before SpikeArrest allows short traffic spikes to consume a partner's daily call allocation before smoothing occurs. Meanwhile, SpikeArrest uses a non-replicated, best-effort cache that can lose counts on cache restarts, making it unsuitable for accurate daily limits.ResponseCache in the TargetEndpoint PreFlow delays cache lookup until late in the proxy execution path, meaning the proxy still executes unnecessary request processing before returning cached data.client_id in the cache key creates duplicate cache entries for each partner even though all partners receive the same catalog data.Quota: Maintains persistent counters in Cassandra for days, weeks, or months, rather than smoothing sub-minute traffic spikes.SpikeArrest: Protects backend capacity over seconds or minutes without storing persistent long-term counts.ProxyEndpoint Request PreFlow and population belongs in the ProxyEndpoint Response PostFlow.No valid Apigee architecture reverses Quota and SpikeArrest in this way; SpikeArrest should always precede Quota, and ResponseCache lookup should occur in the ProxyEndpoint Request PreFlow without client_id when responses are identical across clients.
This configuration combines Apigee's three core traffic management policies—SpikeArrest, synchronous Quota, and ResponseCache—in the recommended ProxyEndpoint flow sequence to shield a Cloud Run backend while preserving partner SLA accuracy and high cache efficiency.
SpikeArrest at the very start of the ProxyEndpoint Request PreFlow acts as the first line of defense, smoothing sudden traffic surges from buggy clients before downstream policies run.Quota policy after VerifyAPIKey and SpikeArrest prevents bursts from prematurely exhausting partner quotas while guaranteeing exact, cross-region call counting in Cassandra.ResponseCache lookup in the ProxyEndpoint Request PreFlow returns cached catalog data early, while populating the cache as late as possible in the ProxyEndpoint Response PostFlow caches the final mediated payload. Using <SkipCachePopulation>response.status.code != 200 or request.verb != "GET"</SkipCachePopulation> ensures errors and write operations are never cached, and omitting client_id shares a single cache entry across all partners.429 Too Many Requests bursts from skewing long-term SLA accounting.Ordering SpikeArrest, VerifyAPIKey, synchronous Quota, and ResponseCache in this exact pattern follows Apigee's design guidelines for high availability, strict SLA compliance, and optimal cache hit rates.
This setup relies solely on SpikeArrest for both short-term traffic smoothing and 24-hour partner quotas, while populating ResponseCache in the Response PreFlow with per-client cache keys.
SpikeArrest is built for short intervals (seconds or minutes) and relies on an un-replicated best-effort Redis cache. Counts can be lost if cache servers restart, and smoothing is not coordinated across Message Processors, so it cannot accurately enforce contractual daily partner limits.ProxyEndpoint Response PreFlow happens before response transformation or mediation policies finish. Caching in Response PostFlow captures the final mediated payload and avoids repeating transformation work on cache hits.<SkipCachePopulation>, the cache could store non-200 error responses or non-GET operations.client_id to the cache key wastes memory by storing identical catalog payloads separately for every partner.SpikeArrest: Returns HTTP 429 during sudden surges, but does not track per-developer or per-app SLAs.request.querystring without client_id maximizes cache hit ratios for shared read-only endpoints.You only needed approximate short-term rate smoothing without contractual daily quotas, and each partner received unique, client-specific data requiring client_id in the cache key.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated