This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
A query returns 40 items and charges 380 request units (RU). Its metrics report a retrieved document count of 92,000, an output document count of 40, and an index lookup time near zero. What does this pattern indicate?
The query fans out across every physical partition, so adding the partition key to the filter is the remediation.
The container's provisioned throughput is too low for the query, so raising the throughput reduces the charge.
At least one part of the filter couldn't use the index, so the engine loaded documents and evaluated the filter against each one.
Two containers each report 8 percent of requests returning a 429 status code. In the first, every partition key range sits near 95 percent normalized RU consumption. In the second, one range sits at 100 percent while the rest stay below 25 percent. What's the correct remediation for the second container?
Investigate the partition key, because a single physical partition is absorbing a disproportionate share of the requests.
Raise the provisioned throughput on the container, because 8 percent of requests exceeding capacity is above the healthy range.
Take no action, because the SDK retries rate-limited requests automatically and the application never observes the error.
Index metrics for a query list /status/? under Potential Single Indexes with an index impact score of High. Almost every item in the container has the same value for status. What conclusion is correct?
/status/?
status
The path is guaranteed to reduce the query's request charge, because the impact score already accounts for how many items match the filter.
The path is a hypothesis worth testing, because a High score reflects the query shape rather than the selectivity of the data.
The path is already in use, so no policy change is needed and the metrics are reporting it for completeness.
A product container is partitioned on /categoryId and has four physical partitions. Which filter forces the query to check every physical partition's index?
product
/categoryId
WHERE c.categoryId = 'AE48F0AA-4F65-4734-A4CF-D48B8F82267F' AND c.price > 1000
WHERE c.categoryId IN ('AE48F0AA-4F65-4734-A4CF-D48B8F82267F', '75BF1ACB-168D-469C-9AA3-1FD26BB4EA4C')
WHERE c.categoryId > 'AE48F0AA-4F65-4734-A4CF-D48B8F82267F'
A developer wants to prove that rewriting a query reduced its cost. Which measurement gives a reliable before-and-after comparison?
The request charge, summed across every page of results, with the page size held constant between runs.
The wall-clock time the call takes to return, averaged over several executions.
The normalized RU consumption percentage reported for the container during the test.
You must answer all questions before checking your work.
Was this page helpful?
Need help with this topic?
Want to try using Ask Learn to clarify or guide you through this topic?