Practice Test 4 | Google Cloud Certified Professional Data Engineer | Dumps | Mock Test
You want to migrate your on-premise NoSQL database to Google Cloud Bigtable. After migration of your complete data, you are not getting the desired performance as per Bigtable standards. What could be the reasons for the performance issue? Select all that apply.
A. The table schema is not designed correctly.
B. Give some time to Bigtable to balance your data.
C. Adding nodes to a cluster and checking cluster performance after 1 hour.
D. Cloud Bigtable instance uses HDD disks.
E. Cloud Bigtable instance is a production instance.
Correct answer is A,B and D.
Option A is correct. For getting good performance from Cloud Bigtable, it is important to design table schema efficiently to distribute reads and writes and avoid hot spotting. Selection of row_key is of very important and crucial for performance.
Option B is correct. If you immediately test Cloud Bigtable after loading of data, performance will be not as per standard. Cloud Bigtable won’t be able to balance your data. It needs time to learn pattern from your data and create large enough shards for effectively use all of the nodes in your cluster.
Option C is incorrect. After addition of nodes it takes approximately 20 minutes before you see performance improvement. In the option user query the data after 1 hour, so the performance improvement must be done by then.
Option D is correct. SDD disk has better performance that HDD disk. It is recommended to use SDD disk for performance sensitive applications.
Option E is incorrect. As Cloud Bigtable instance is already a production instance, there is no scope of further improvement by changing the instance type. Also, development instance has less performance than production instance. A development instance has a performance equivalent to one single node cluster.
Comments are closed, but trackbacks and pingbacks are open.