Load-test your database

Runs other people can check

Each row is a run published from Stroppy Cloud, with the workload, the database and its version, and the machines it ran on. Filter by any of them; a row opens the full report with every metric of that run.

Loading the leaderboard…

Which databases really speak the protocol they claim

Each one measured check by check against the original: PostgreSQL, MySQL, MongoDB, Elasticsearch.

Works the same
Differs
Refused
Breaks
Not measured

PostgreSQL	Score	Connection and auth	Query flows	Errors, cancel and pipeline	Session and notices	Numbers	Text and binary	Date and time	Structured	Special	Type contract	Bulk load
						Types
Referencepostgres-13 to postgres-18	the original	Works the same, 86 of 86 checks matched	Works the same, 72 of 72 checks matched	Works the same, 60 of 60 checks matched	Works the same, 36 of 36 checks matched	Works the same, 96 of 96 checks matched	Works the same, 66 of 66 checks matched	Works the same, 96 of 96 checks matched	Works the same, 522 of 522 checks matched	Works the same, 246 of 246 checks matched	Works the same, 2004 of 2004 checks matched	Works the same, 4380 of 4380 checks matched
alloydb-omni	98.9	Refused, 11 of 12 checks matched	Works the same, 12 of 12 checks matched	Works the same, 10 of 10 checks matched	Works the same, 6 of 6 checks matched	Works the same, 16 of 16 checks matched	Works the same, 11 of 11 checks matched	Works the same, 16 of 16 checks matched	Works the same, 87 of 87 checks matched	Works the same, 41 of 41 checks matched	Works the same, 334 of 334 checks matched	Works the same, 735 of 735 checks matched
cloudberry	98.7	Refused, 11 of 12 checks matched	Works the same, 12 of 12 checks matched	Works the same, 10 of 10 checks matched	Works the same, 6 of 6 checks matched	Works the same, 16 of 16 checks matched	Works the same, 11 of 11 checks matched	Works the same, 16 of 16 checks matched	Works the same, 87 of 87 checks matched	Differs, 40 of 41 checks matched	Works the same, 332 of 332 checks matched	Differs, 722 of 727 checks matched
yugabytedb	97.7	Refused, 11 of 12 checks matched	Works the same, 12 of 12 checks matched	Works the same, 10 of 10 checks matched	Refused, 5 of 6 checks matched	Works the same, 16 of 16 checks matched	Works the same, 11 of 11 checks matched	Works the same, 16 of 16 checks matched	Differs, 86 of 87 checks matched	Works the same, 41 of 41 checks matched	Works the same, 334 of 334 checks matched	Differs, 732 of 733 checks matched
pgbouncer	91.3	Differs, 9 of 12 checks matched	Breaks, 11 of 12 checks matched	Breaks, 8 of 10 checks matched	Differs, 4 of 6 checks matched	Works the same, 16 of 16 checks matched	Works the same, 11 of 11 checks matched	Works the same, 16 of 16 checks matched	Works the same, 87 of 87 checks matched	Works the same, 41 of 41 checks matched	Works the same, 334 of 334 checks matched	Differs, 736 of 737 checks matched
cockroachdb	68.2	Breaks, 7 of 15 checks matched	Differs, 11 of 12 checks matched	Refused, 9 of 10 checks matched	Differs, 1 of 6 checks matched	Differs, 13 of 16 checks matched	Differs, 8 of 11 checks matched	Differs, 9 of 16 checks matched	Breaks, 24 of 81 checks matched	Differs, 7 of 41 checks matched	Differs, 87 of 89 checks matched	Breaks, 281 of 332 checks matched

Open the compatibility mapThe five best of 20 PostgreSQL engines. Dataset 45a7361f915b.

Prefer the terminal?

Stroppy CLI is open source and runs against any database it has a driver for: one Go binary with the workloads, the drivers and the report inside. Hover a note to see which part of the run it’s about.

curl -fsSL https://github.com/stroppy-io/stroppy/releases/latest/download/stroppy_linux_amd64.tar.gz | tar -xz
sudo install stroppy /usr/local/bin/

It’s one file, about 40 MB, and the archive has nothing else in it. Docker and building from source are covered in the introduction.

~/bench

$ stroppy run tpcb/tx -d pg -D url=postgres://localhost/bench --executor constant-vus --vus 8 --duration 30s
19:08:49.103  INFO  End of 'create_schema' step (took 24ms)
19:08:49.236  INFO  End of 'load_data' step (took 132ms)
19:08:49.553  INFO  End of 'analyze' step (took 262ms)

=== bench summary ===
  transactions_total                       5603.000
  failed_iterations_total                  0.000
  tx_total_duration
    count=5603 avg=42.767 p(50)~=50.000 p(90)~=100.000 p(95)~=100.000 p(99)~=250.000
  tx_commits_total                         5603.000
  tps                                      186.530
  queries_per_second                       932.648
  measurement_seconds                      30.038

Real output from Stroppy 6.1.1 against PostgreSQL 17 in Docker, trimmed to the lines the notes talk about.

What you ask for

What you get back

How fast the client is, and why it matters

Every number a load generator reports includes the load generator. If the client is slow, the database waits and you end up measuring the tool. stroppy baseline shows how much room it leaves on your machine.

go-tpc → StroppyTPC-C on OrioleDB, one hour, 1000 warehouses, 100 VUs, Stroppy 5.7.2. How we found it

× go-tpc

go-tpc → tpcc/procs

v5 on k6 → v6Medians of five runs on one AMD EPYC Genoa VM. Big on the noop driver, small on a real database, which is how it should be. Why we dropped k6

× v5

TPC-B ceiling, 1 VU

TPC-B ceiling, 24 VUs

TPC-C on PostgreSQL, 1 VU

TPC-C on PostgreSQL, 16 VUs

Database → Stroppy ceilingSame VM, 16 VUs. Different workloads; the gap is the room the client leaves.

× PostgreSQL

PostgreSQL TPC-C → noop TPC-B

Don’t feel like setting up the machines?

Stroppy Cloud does that part. You pick a database and a workload; it brings up the stand, tunes it, runs the test and gives you a report with queries per second, latency and the moment errors began. Run it again next week and the two sit side by side.

Create account How it works