https://venicedb.org logo
Join Slack
Powered by
# performance-testing
  • e

    Enrico Olivelli

    05/26/2023, 6:41 AM
    Can you please share here a list of the most important metrics on the read path with the regular client ? Something like: • latency on the server • latency on the router
  • e

    Enrico Olivelli

    05/26/2023, 6:42 AM
    I am trying to find the best metrics, but I may be using wrong data
  • e

    Enrico Olivelli

    05/26/2023, 9:20 AM
    Today I have been able to set up some basic Grafana dashboards. These are my numbers so far: • 1 server • 1 router • 1 client, 10 threads, sharing one instance of AvroGenericStoreClient • 1 store • the client accesses only the router using HTTP (no TLS) • same network for all the components • 2.2 ms p50 on the client, 0.6 ms p50 on the router, 0.1ms p50 on the server • 3ms p95 on the client • 22.4 ms p99 on the client, 1.5ms p99 on the router, 0.5ms p99 on the server The difference between the client and the router is huge. I guess that it is due to the AVRO deserialisation. But I would like to understand how can I achieve better numbers. Which are the tunables on the client ? ◦ number of connections ? ◦ turning on HHTP 2.0 ? ◦ do not deserialise ? ◦ should I NOT share the client with multiple threads ? (it looks like we have a great async API, so I assume that having multiple inflight requests is supported, as far as we can open multiple connections or leverage http2 multiplexing)
  • e

    Enrico Olivelli

    05/26/2023, 10:04 AM
    I have run a new test with 3 routers and 3 servers and replication = 3, but it seems that the results are worse on the client 🤔
  • f

    Felix GV

    05/26/2023, 10:05 AM
    The client-side numbers sure look a bit high.
  • f

    Felix GV

    05/26/2023, 10:06 AM
    You could try http2, to see if it’s due to connection starvation
  • e

    Enrico Olivelli

    05/26/2023, 10:34 AM
    how do I enable it ?
  • f

    Felix GV

    05/26/2023, 11:15 AM
    @Gaojie Liu, could you help out with this ^
  • e

    Enrico Olivelli

    05/26/2023, 12:45 PM
    I did more experiments: • the network is not a problem (I see the same results both from a client pod and from the Router pod, also connecting to localhost, instead than passing thru the K8S Service) • Avro is not a problem, there is a “blackhole avro deserializer” option that drops that part I believe that it is more about the number of connections per router. I get better results with only 1 thread in the client
  • s

    Slackbot

    05/26/2023, 12:48 PM
    This message was deleted.
    e
    f
    z
    • 4
    • 43
  • s

    Slackbot

    05/26/2023, 1:24 PM
    This message was deleted.
    f
    e
    • 3
    • 2
  • s

    Slackbot

    05/26/2023, 1:45 PM
    This message was deleted.
    f
    • 2
    • 2
  • s

    Slackbot

    05/26/2023, 3:47 PM
    This message was deleted.
    🚀 3
    f
    e
    • 3
    • 2
  • e

    Enrico Olivelli

    05/26/2023, 3:47 PM
    I will be back on this on Monday
  • z

    Zac Policzer

    05/26/2023, 3:47 PM
    Nice
  • z

    Zac Policzer

    05/26/2023, 3:48 PM
    Out of curiosity, what latency targets are datastax looking for?
  • e

    Enrico Olivelli

    05/27/2023, 6:04 AM
    Frankly speaking, the goal si to be able to say that Venice is generally better than Redis on reads 😉 That said, achieving <1ms p50 and <10ms p99 from a client on the same “network” (cloud AZ) as Venice would be great (maybe @Chris Bartholomew / @Dave Fisher could add some details or refine the goal)
  • f

    Felix GV

    05/27/2023, 9:31 AM
    I think competing with Redis is possible but needs further tuning, such as using the PlainTable mode.
  • z

    Zac Policzer

    05/27/2023, 3:00 PM
    I look forward to sneering at that "fastest database" Redis billboard along the 101 here 😂🤣
    😁 2
    😂 2
  • s

    Slackbot

    05/27/2023, 4:09 PM
    This message was deleted.
    f
    e
    • 3
    • 6
  • s

    Slackbot

    05/27/2023, 4:10 PM
    This message was deleted.
    f
    e
    • 3
    • 14
  • s

    Slackbot

    05/29/2023, 11:05 AM
    This message was deleted.
    e
    • 2
    • 2