Bringing together the Apache Cassandra experts from the community and DataStax.

Want to learn? Have a question? Want to share your expertise? You are in the right place!

Not sure where to begin? Getting Started

 

question

praveenkg avatar image
praveenkg asked Erick Ramirez answered

How can we optimize a single-partition read to be below 100ms?

I am working on cassandra application having a table modelled based on item number as partition key. Table has roughly 25K records. I am querying the table using Java application and i see for the one item, it takes more than 100ms to retrieve the data, I enabled the tracing in java and below are the results

'Execute CQL3 prepared query' to /<host>:0 took 1712μs
2022-02-10 10:16:03.414 INFO 33444 --- [ s0-io-5] c.d.o.d.i.core.tracker.RequestLogger : [s0|1449246075][Node(endPoint=/<hoost2>:9042, hostId=6100a2d7-cce9-409f-ac88-a7257daa027c, hashCode=b7f1344)] Success (365 ms) [1 values] SELECT * FROM system_traces.events WHERE session_id = ? [v0=59722650-8a2c-11ec-8a52-e97691b34e4a]
303 - /<hoost1>:0 - Reading data from [/<hoost1>]
303 - /<hoost1>:0 - Reading digests from [/<hoost2>]
401 - /<hoost1>:0 - Skipped 0/0 non-slice-intersecting sstables, included 0 due to tombstones
401 - /<hoost1>:0 - Merged data from memtables and 0 sstables during single-partition query on inv_by_item_loc for key '6202620010009'
401 - /<hoost1>:0 - Sending READS.SINGLE_READ message to /<hoost2>, size=420 bytes
401 - /<hoost1>:0 - Read 3 live rows and 0 tombstone ones
1500 - /<hoost1>:0 - READS.SINGLE_READ message received from /<hoost2>
0 - /<hoost2>:0 - READS.SINGLE_READ message received from /<hoost1>
1500 - /<hoost1>:0 - Processing response from /<hoost2>

is there any optimization we can do bring down the response time below 100ms

performance
10 |1000

Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.

1 Answer

Erick Ramirez avatar image
Erick Ramirez answered

There are a lot of factors which would affect the performance of reads on your cluster including:

  • cluster size and topology
  • data model
  • access patterns
  • hardware configuration
  • cluster load
  • etc

This isn't something we can answer in a Q&A forum. I suggest you log a ticket with DataStax Support so one of our engineers can request the necessary diagnostic data to assist you with the investigation. Cheers!

Share
10 |1000

Up to 8 attachments (including images) can be used with a maximum of 1.0 MiB each and 10.0 MiB total.