daml/ledger-service/http-json-perf
2021-09-07 23:42:30 +01:00
..
daml [JSON-API] Perf gatling MultiUserQueryScenario (#10422) 2021-08-16 17:01:27 +00:00
release Async logging for http-json-perf tests and publish logs as a pipeline artifact (#10803) 2021-09-07 23:42:30 +01:00
src/main [JSON-API] Use the token from incoming requests to update the package list (#10602) 2021-09-03 10:12:01 +00:00
BUILD.bazel Upgrade Scalatest to v3.2.9. (#10576) 2021-08-12 23:19:35 +00:00
README.md [JSON-API] key_hash field to speed up fetchByKey queries (#10631) 2021-08-23 18:15:25 +01:00

1. Gatling Scenarios

1.1. Prerequisites

All current Gatling scenarios require quickstart-model.dar with IOU example. You can build one using:

bazel build //docs:quickstart-model
ls "${PWD}/bazel-bin/docs/quickstart-model.dar"

1.2. List of Scenarios

Gatling scenarios extend from io.gatling.core.scenario.Simulation:

  • com.daml.http.perf.scenario.CreateCommand
  • com.daml.http.perf.scenario.ExerciseCommand
  • com.daml.http.perf.scenario.CreateAndExerciseCommand
  • com.daml.http.perf.scenario.AsyncQueryConstantAcs
  • com.daml.http.perf.scenario.SyncQueryConstantAcs
  • com.daml.http.perf.scenario.SyncQueryNewAcs
  • com.daml.http.perf.scenario.SyncQueryVariableAcs
  • com.daml.http.perf.scenario.OracleMultiUserQueryScenario

2. Running Gatling Scenarios from Bazel

2.1. Help

$ bazel run //ledger-service/http-json-perf:http-json-perf-binary -- --help

2.2. Example

$ bazel run //ledger-service/http-json-perf:http-json-perf-binary -- \
--scenario=com.daml.http.perf.scenario.CreateCommand \
--dars="${PWD}/bazel-bin/docs/quickstart-model.dar" \
--reports-dir=/home/leos/tmp/results/ \
--jwt="eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJodHRwczovL2RhbWwuY29tL2xlZGdlci1hcGkiOnsibGVkZ2VySWQiOiJNeUxlZGdlciIsImFwcGxpY2F0aW9uSWQiOiJmb29iYXIiLCJhY3RBcyI6WyJBbGljZSJdfX0.VdDI96mw5hrfM5ZNxLyetSVwcD7XtLT4dIdHIOa9lcU"

2.3 Running OracleMultiUserQueryScenario

We use an external docker oracle vm, so we want to retain the data between runs to specifically focus on testing query performance. use RETAIN_DATA and USE_DEFAULT_USER env vars to use a static user(ORACLE_USER) and preserve data. This scenario uses a single template KeyedIou defined in LargeAcs.daml.

We can control a few scenario parameters i.e NUM_RECORDS NUM_QUERIES NUM_READERS NUM_WRITERS via env variables

  1. Populate Cache

USE_DEFAULT_USER=true RETAIN_DATA=true RUN_MODE="populateCache" bazel run //ledger-service/http-json-perf:http-json-perf-binary-ee -- --scenario=com.daml.http.perf.scenario.OracleMultiUserQueryScenario --jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJodHRwczovL2RhbWwuY29tL2xlZGdlci1hcGkiOnsibGVkZ2VySWQiOiJNeUxlZGdlciIsImFwcGxpY2F0aW9uSWQiOiJmb29iYXIiLCJhY3RBcyI6WyJBbGljZSJdfX0.VdDI96mw5hrfM5ZNxLyetSVwcD7XtLT4dIdHIOa9lcU --dars=$PWD/bazel-bin/ledger-service/http-json-perf/LargeAcs.dar --query-store-index oracle

  1. Fetch By Key

Query contracts by the defined key field.


USE_DEFAULT_USER=true RETAIN_DATA=true RUN_MODE="fetchByKey" NUM_QUERIES=100 bazel run //ledger-service/http-json-perf:http-json-perf-binary-ee -- --scenario=com.daml.http.perf.scenario.OracleMultiUserQueryScenario --jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJodHRwczovL2RhbWwuY29tL2xlZGdlci1hcGkiOnsibGVkZ2VySWQiOiJNeUxlZGdlciIsImFwcGxpY2F0aW9uSWQiOiJmb29iYXIiLCJhY3RBcyI6WyJBbGljZSJdfX0.VdDI96mw5hrfM5ZNxLyetSVwcD7XtLT4dIdHIOa9lcU --dars=$PWD/bazel-bin/ledger-service/http-json-perf/LargeAcs.dar --query-store-index oracle

  1. Fetch By Query

Query contracts by a field on the payload which is the id in this case.


USE_DEFAULT_USER=true RETAIN_DATA=true RUN_MODE="fetchByQuery" bazel run //ledger-service/http-json-perf:http-json-perf-binary-ee -- --scenario=com.daml.http.perf.scenario.OracleMultiUserQueryScenario --jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJodHRwczovL2RhbWwuY29tL2xlZGdlci1hcGkiOnsibGVkZ2VySWQiOiJNeUxlZGdlciIsImFwcGxpY2F0aW9uSWQiOiJmb29iYXIiLCJhY3RBcyI6WyJBbGljZSJdfX0.VdDI96mw5hrfM5ZNxLyetSVwcD7XtLT4dIdHIOa9lcU --dars=$PWD/bazel-bin/ledger-service/http-json-perf/LargeAcs.dar --query-store-index oracle

3. Running Gatling Scenarios Manually

The following instructions tested on Linux but should also work on macOs.

3.1. Install Gatling (open-source load testing solution)

3.2. Create quickstart Daml project

$ daml new quickstart-java --template quickstart-java
$ cd quickstart-java/
$ daml build

3.3. Start Sandbox with quickstart DAR

Ledger ID MyLedger is important, it is currently hardcoded in the com.daml.http.perf.scenario.SimulationConfig. See aliceJwt.

$ daml sandbox --ledgerid MyLedger ./.daml/dist/quickstart-0.0.1.dar

3.4. Start JSON API

daml json-api  --ledger-host=localhost --ledger-port=6865 --http-port=7575 --package-reload-interval 5h --allow-insecure-tokens

3.5. Run Gatling scenario

$ <GATLING_HOME>/bin/gatling.sh --simulations-folder=<DAML_PROJECT_HOME>/ledger-service/http-json-perf/src/main/scala/com/daml/http/perf/scenario --simulation=com.daml.http.perf.scenario.CreateCommand

Where:

  • <GATLING_HOME> -- path to the Gatling directory
  • <DAML_PROJECT_HOME> -- path to the Daml Repository on the local disk
  • --simulation=com.daml.http.perf.scenario.CreateCommand -- full class name of the scenario from the --simulations-folder