Built-in Workloads

Stroppy v6 compiles workload implementations and their assets into one binary. No workspace generation is required.

stroppy run simple
stroppy run tpcb/tx
stroppy run tpcb/procs
stroppy run tpcc/tx
stroppy run tpcc/procs
stroppy run tpch/tx
stroppy run tpcds
stroppy run ./queries.sql

Catalog

Workload	Purpose	Drivers
simple	Small create/load/read/drop smoke	SQL drivers with row results
tpcb/tx	TPC-B client-side transaction	PostgreSQL, MySQL, Picodata, YDB
tpcb/procs	TPC-B stored procedure	PostgreSQL, MySQL
tpcc/tx	TPC-C five-transaction mix	PostgreSQL, MySQL, Picodata, YDB
tpcc/procs	TPC-C stored procedures	PostgreSQL, MySQL
tpch/tx	TPC-H load plus q1–q22	PostgreSQL, MySQL, Picodata, YDB
tpcds	TPC-DS load plus query suite	PostgreSQL, MySQL, YDB; Picodata load-only
execute_sql	SQL file or inline SQL	SQL query drivers

Discover exact schemas and embedded assets:

stroppy probe
stroppy probe -o json
stroppy run tpcc/tx --help

Shared executors

All registered workloads accept:

Flag	Default	Meaning
--executor	shared-iterations	shared-iterations or constant-vus.
--vus	1	Virtual users.
--iterations	1	Total shared iterations.
--duration	0s	Constant-VU duration.
--query-timeout	0s	Per-statement deadline; zero disables it.

Power/fixed-work run:

stroppy run tpcb/tx \
  --executor shared-iterations --vus 4 --iterations 100

Throughput run:

stroppy run tpcc/tx \
  --executor constant-vus --vus 32 --duration 5m

Two-pass load and measure

Most database workloads run setup before their executor. Split loading from measurement when load time must not affect throughput interpretation.

# Prepare data.
stroppy run tpcc/tx -d pg --scale-factor 50 --load-workers 16 \
  --no-steps workload

# Measure existing data.
stroppy run tpcc/tx -d pg --scale-factor 50 \
  --executor constant-vus --vus 64 --duration 1h \
  --steps workload

Step filters apply to both passes. Ensure both commands use the same database, scale, SQL variant, and workload-specific range settings.

simple

simple demonstrates the native workload lifecycle:

1. drop stroppy_demo;
2. create three-column table;
3. load 100 deterministic rows through a typed insert request;
4. validate row count and perform lookups;
5. drop table during teardown.

stroppy run simple -d pg

Steps: drop_schema, create_schema, load_data, workload.

The workload validates returned rows, so it is not a Noop-driver overhead test. Use stroppy baseline for Noop and wire ceilings.

tpcb/tx and tpcb/procs

TPC-B models a banking transaction: update account, read balance, update teller and branch, then append history.

stroppy run tpcb/tx -d pg --scale-factor 10 \
  --executor constant-vus --vus 32 --duration 5m

stroppy run tpcb/procs -d mysql --iterations 100

Typed workload flags:

Flag	Default	Meaning
--scale-factor	1	Branch count; 10 tellers and 100,000 accounts per branch.
--load-workers	1	Workers used for each table load.
--retry-attempts	3	Maximum transaction attempts.
--tx-isolation	driver-derived	Isolation override.
--sql-file	dialect asset	SQL override.

Default isolation: read_committed on PostgreSQL/MySQL, none on Picodata, serializable on YDB.

Setup order:

drop_schema
create_schema
create_procedures       # procs only
load_data
create_indexes
create_foreign_keys
analyze
workload

See TPC-B.

tpcc/tx and tpcc/procs

TPC-C runs the standard 45/43/4/4/4 New-Order, Payment, Order-Status, Delivery, and Stock-Level mix.

stroppy run tpcc/tx -d pg --scale-factor 10 \
  --executor constant-vus --vus 32 --duration 5m

stroppy run tpcc/procs -d mysql --pacing=true --iterations 100

Typed workload flags:

Flag	Default	Meaning
--scale-factor	1	Warehouse count; legacy alias WAREHOUSES.
--warehouse-start	1	First warehouse ID for distributed loading.
--load-items	contextual	Load global item table; defaults true only at warehouse 1.
--load-workers	1	Workers per table load.
--pacing	false	Apply TPC-C keying and think times.
--retry-attempts	3	Maximum transaction attempts.
--pg-unlogged	false	Opt into PostgreSQL unlogged load.
--tx-isolation	driver-derived	Isolation override.
--sql-file	dialect asset	SQL override.

Default isolation: repeatable_read on PostgreSQL/MySQL, none on Picodata, serializable on YDB.

Setup order:

drop_schema
create_schema
create_procedures       # procs only
set_unlogged            # opt-in PostgreSQL
load_data
create_indexes
set_logged              # when set_unlogged ran
create_foreign_keys
analyze
validate_population
workload

A population-validation failure is structural and exits nonzero. Both variants emit per-transaction count, mix, throughput, and p50/p90/p95/p99 summaries. Paced runs additionally report response-time/mix verdicts and steady-state assessment.

See TPC-C.

tpch/tx

TPC-H loads eight canonical dbgen tables and executes q1–q22 with pinned parameters.

stroppy run tpch/tx -d pg --scale-factor 0.01
stroppy run tpch/tx -d pico --scale-factor 0.01

Typed workload flags:

Flag	Default	Meaning
--scale-factor	1	Positive fractional row scale.
--load-workers	0	Load workers; zero resolves to one.
--pg-unlogged	false	Opt into PostgreSQL unlogged load.
--ydb-store-mode	column	column or row.
--sql-file	dialect asset	SQL override.

Setup order:

drop_schema
create_schema
set_unlogged            # opt-in PostgreSQL
load_data
create_indexes
set_logged              # when set_unlogged ran
analyze
validate_answers
workload

o_totalprice is computed during canonical generation; no post-load finalize step exists. SF=1 answer comparison is PostgreSQL-only and diagnostic.

See TPC-H.

tpcds

TPC-DS loads all 24 canonical dsdgen tables. PostgreSQL, MySQL, and YDB execute the baked query suite; Picodata supports loading only.

stroppy run tpcds -d pg --scale-factor 0.1
stroppy run tpcds -d ydb --scale-factor 0.1
stroppy run tpcds -d pico --scale-factor 0.1 --no-steps workload

Typed workload flags:

Flag	Default	Meaning
--scale-factor	1	Positive fractional row scale.
--load-workers	0	Workers per table load.
--pg-unlogged	false	Opt into PostgreSQL unlogged load.
--ydb-store-mode	column	column or row.
--streams	1	Number of generated query streams.
--query-stream	unset	Explicit generated stream number.
--query-seed	19620718	Generated-stream seed.
--schema-file	dialect asset	Schema override.
--sql-file	dialect asset	Query override.
--validate-force	false	Compare SF=1 references outside SF=1.

YDB accepts only the baked query set; generated streams are rejected. Answer comparison is PostgreSQL/MySQL-only and diagnostic. Several static dimensions remain large even at fractional scale.

Setup order:

drop_schema
create_schema
set_unlogged            # opt-in PostgreSQL
load_data
create_indexes
set_logged              # when set_unlogged ran
analyze
validate_answers         # baked PG/MySQL path
workload

See TPC-DS.

execute_sql

A .sql first positional or inline statement selects execute_sql:

stroppy run ./queries.sql -d pg
stroppy run "select 1" -d pg

Workload-specific flags:

Flag	Default	Meaning
--sql-file	unset	SQL file path.
--sql-body	unset	Inline SQL body.

Shared executor and query-timeout flags also apply. Named queries in a flat file run in order. Nonfatal query errors are counted and the remaining query set continues.

Environment compatibility

Uppercase environment names and -e remain supported, but direct flags are preferred:

stroppy run tpcc/tx -e warehouses=10 -e load_workers=8

Typed source precedence and config mapping are documented in Configuration Files.