Timestream benchmark code

Hello !

I would like to reproduce the results of Amazon Timestream is finally released: these are our first impressions

Did you put the code for this in open access somewhere ?

It would really help me out, thanks.

Hi @Hadrien-Cornier,

That code sits within a benchmark project with some other tooling, which unfortunately right now isn’t openly accessible. Would have to check if we could do that.

May I ask what you want to do with it? are you only interested in the aws timestream part?

Hello, thanks for your quick answer.
I’m comparing different databases for a business project and I’ve seen that you claim to be much better than AWS Timestream but I would like to verify for myself if your numbers are accurate. It’s difficult to reproduce your results because there is not enough details on your blogpost so the code would be very useful.

sure, I understand

maybe lets give you some insights on the benchmarks:

its all a python project that is prepared to run distributed in a kubernetes cluster.
we were using amazons boto3 (TimestreamWrite — Boto3 Docs 1.17.33 documentation) with

  • read_timeout=20
  • max_pool_connections=5000
  • retries={‘max_attempts’: 10}

20 pods / clients were running parallel during the insert tests
10 clients were running the query load

Dear Hadrien,

thank you for writing in.

We did not release it yet, but planned to do so. I will check if we can do that soon and will let you know immediately after.

With kind regards,
Andreas.