OpenSearch Benchmark quickstart
This tutorial outlines how to quickly install OpenSearch Benchmark and run your first OpenSearch Benchmark workload.
Prerequisites
To perform the Quickstart steps, you’ll need to fulfill the following prerequisites:
- A currently active OpenSearch cluster. For instructions on how to create an OpenSearch cluster, see Creating a cluster.
- Git 2.3 or greater.
- Python 3.8 or later
Set up an OpenSearch cluster
If you don’t already have an active OpenSearch cluster, you can launch a new OpenSearch cluster to use with OpenSerch Benchmark.
- Using Docker Compose. For instructions on how to use Docker Compose, see OpenSearch Quickstart.
- Using Tar. For instructions on how to install OpenSearch with Tar, see Installing OpenSearch > Tarball.
OpenSearch Benchmark has not been tested with the Window’s distribution of OpenSearch.
After installation, you can verify OpenSearch is running by going to localhost:9200. If you’re running your cluster with the Security plugin enabled, OpenSearch will expect SSL connections with the username “admin” and password “admin”. However, since the localhost address is not a unique public address, no certificate authority will issue an SSL certificate for it, so certificate checking will need to be disabled using the -k option.
Use the following command to verify OpenSearch is running with SSL certificate checks disabled:
curl -k -u admin:<custom-admin-password> https://localhost:9200 # the "-k" option skips SSL certificate checks
{
"name" : "147ddae31bf8.opensearch.org",
"cluster_name" : "opensearch",
"cluster_uuid" : "n10q2RirTIuhEJCiKMkpzw",
"version" : {
"distribution" : "opensearch",
"number" : "2.10.0",
"build_type" : "tar",
"build_hash" : "eee49cb340edc6c4d489bcd9324dda571fc8dc03",
"build_date" : "2023-09-20T23:54:29.889267151Z",
"build_snapshot" : false,
"lucene_version" : "9.7.0",
"minimum_wire_compatibility_version" : "7.10.0",
"minimum_index_compatibility_version" : "7.0.0"
},
"tagline" : "The OpenSearch Project: https://opensearch.org/"
}
With your cluster running, you can now install OpenSearch Benchmark.
Installing OpenSearch Benchmark
To install OpenSearch Benchmark with Docker, see Installing OpenSearch Benchmark > Installing with Docker.
To install OpenSearch Benchmark from PyPi, enter the following pip command:
pip3 install opensearch-benchmark
After the installation completes, verify that OpenSearch Benchmark is running by entering the following command:
opensearch-benchmark --help
If successful, OpenSearch returns the following response:
$ opensearch-benchmark --help
usage: opensearch-benchmark [-h] [--version] {execute-test,list,info,create-workload,generate,compare,download,install,start,stop} ...
____ _____ __ ____ __ __
/ __ \____ ___ ____ / ___/___ ____ ___________/ /_ / __ )___ ____ _____/ /_ ____ ___ ____ ______/ /__
/ / / / __ \/ _ \/ __ \\__ \/ _ \/ __ `/ ___/ ___/ __ \ / __ / _ \/ __ \/ ___/ __ \/ __ `__ \/ __ `/ ___/ //_/
/ /_/ / /_/ / __/ / / /__/ / __/ /_/ / / / /__/ / / / / /_/ / __/ / / / /__/ / / / / / / / / /_/ / / / ,<
\____/ .___/\___/_/ /_/____/\___/\__,_/_/ \___/_/ /_/ /_____/\___/_/ /_/\___/_/ /_/_/ /_/ /_/\__,_/_/ /_/|_|
/_/
A benchmarking tool for OpenSearch
optional arguments:
-h, --help show this help message and exit
--version show program's version number and exit
subcommands:
{execute-test,list,info,create-workload,generate,compare,download,install,start,stop}
execute-test Run a benchmark
list List configuration options
info Show info about a workload
create-workload Create a Benchmark workload from existing data
generate Generate artifacts
compare Compare two test_executions
download Downloads an artifact
install Installs an OpenSearch node locally
start Starts an OpenSearch node locally
stop Stops an OpenSearch node locally
Find out more about Benchmark at https://opensearch.org/docs
Running your first benchmark
You can now run your first benchmark. The following benchmark uses the percolator workload.
Understanding workload command flags
Benchmarks are run using the execute-test command with the following command flags:
For additional execute_test command flags, see the execute-test reference. Some commonly used options are --workload-params, --exclude-tasks, and --include-tasks.
--pipeline=benchmark-only: Informs OSB that users wants to provide their own OpenSearch cluster.workload=percolator: The name of workload used by OpenSearch Benchmark.--target-host="<OpenSearch Cluster Endpoint>": Indicates the target cluster or host that will be benchmarked. Enter the endpoint of your OpenSearch cluster here.--client-options="basic_auth_user:'<Basic Auth Username>',basic_auth_password:'<Basic Auth Password>'": The username and password for your OpenSearch cluster.--test-mode: Allows a user to run the workload without running it for the entire duration. When this flag is present, Benchmark runs the first thousand operations of each task in the workload. This is only meant for sanity checks—the metrics produced are meaningless.
The --distribution-version, which indicates which OpenSearch version Benchmark will use when provisioning. When run, the execute-test command will parse the correct distribution version when it connects to the OpenSearch cluster.
Running the workload
To run the percolator workload with OpenSearch Benchmark, use the following execute-test command:
opensearch-benchmark execute-test --pipeline=benchmark-only --workload=percolator --target-host=https://localhost:9200 --client-options=basic_auth_user:admin,basic_auth_password:admin,verify_certs:false --test-mode
When the execute_test command runs, all tasks and operations in the percolator workload run sequentially.
Validating the test
After an OpenSearch Benchmark test runs, take the following steps to verify that it has run properly:
- Note the number of documents in the OpenSearch or OpenSearch Dashboards index that you plan to run the benchmark against.
- In the results returned by OpenSearch Benchmark, compare the
workload.jsonfile for your specific workload and verify that the document count matches the number of documents. For example, based on the percolatorworkload.jsonfile, you should expect to see2000000documents in your cluster.
Understanding the results
OpenSearch Benchmark returns the following response once the benchmark completes:
------------------------------------------------------
_______ __ _____
/ ____(_)___ ____ _/ / / ___/_________ ________
/ /_ / / __ \/ __ `/ / \__ \/ ___/ __ \/ ___/ _ \
/ __/ / / / / / /_/ / / ___/ / /__/ /_/ / / / __/
/_/ /_/_/ /_/\__,_/_/ /____/\___/\____/_/ \___/
------------------------------------------------------
| Metric | Task | Value | Unit |
|---------------------------------------------------------------:|-------------------------------------------:|------------:|-------:|
| Cumulative indexing time of primary shards | | 0.02655 | min |
| Min cumulative indexing time across primary shards | | 0 | min |
| Median cumulative indexing time across primary shards | | 0.00176667 | min |
| Max cumulative indexing time across primary shards | | 0.0140333 | min |
| Cumulative indexing throttle time of primary shards | | 0 | min |
| Min cumulative indexing throttle time across primary shards | | 0 | min |
| Median cumulative indexing throttle time across primary shards | | 0 | min |
| Max cumulative indexing throttle time across primary shards | | 0 | min |
| Cumulative merge time of primary shards | | 0.0102333 | min |
| Cumulative merge count of primary shards | | 3 | |
| Min cumulative merge time across primary shards | | 0 | min |
| Median cumulative merge time across primary shards | | 0 | min |
| Max cumulative merge time across primary shards | | 0.0102333 | min |
| Cumulative merge throttle time of primary shards | | 0 | min |
| Min cumulative merge throttle time across primary shards | | 0 | min |
| Median cumulative merge throttle time across primary shards | | 0 | min |
| Max cumulative merge throttle time across primary shards | | 0 | min |
| Cumulative refresh time of primary shards | | 0.0709333 | min |
| Cumulative refresh count of primary shards | | 118 | |
| Min cumulative refresh time across primary shards | | 0 | min |
| Median cumulative refresh time across primary shards | | 0.00186667 | min |
| Max cumulative refresh time across primary shards | | 0.0511667 | min |
| Cumulative flush time of primary shards | | 0.00963333 | min |
| Cumulative flush count of primary shards | | 4 | |
| Min cumulative flush time across primary shards | | 0 | min |
| Median cumulative flush time across primary shards | | 0 | min |
| Max cumulative flush time across primary shards | | 0.00398333 | min |
| Total Young Gen GC time | | 0 | s |
| Total Young Gen GC count | | 0 | |
| Total Old Gen GC time | | 0 | s |
| Total Old Gen GC count | | 0 | |
| Store size | | 0.000485923 | GB |
| Translog size | | 2.01873e-05 | GB |
| Heap used for segments | | 0 | MB |
| Heap used for doc values | | 0 | MB |
| Heap used for terms | | 0 | MB |
| Heap used for norms | | 0 | MB |
| Heap used for points | | 0 | MB |
| Heap used for stored fields | | 0 | MB |
| Segment count | | 32 | |
| Min Throughput | index | 3008.97 | docs/s |
| Mean Throughput | index | 3008.97 | docs/s |
| Median Throughput | index | 3008.97 | docs/s |
| Max Throughput | index | 3008.97 | docs/s |
| 50th percentile latency | index | 351.059 | ms |
| 100th percentile latency | index | 365.058 | ms |
| 50th percentile service time | index | 351.059 | ms |
| 100th percentile service time | index | 365.058 | ms |
| error rate | index | 0 | % |
| Min Throughput | wait-until-merges-finish | 28.41 | ops/s |
| Mean Throughput | wait-until-merges-finish | 28.41 | ops/s |
| Median Throughput | wait-until-merges-finish | 28.41 | ops/s |
| Max Throughput | wait-until-merges-finish | 28.41 | ops/s |
| 100th percentile latency | wait-until-merges-finish | 34.7088 | ms |
| 100th percentile service time | wait-until-merges-finish | 34.7088 | ms |
| error rate | wait-until-merges-finish | 0 | % |
| Min Throughput | percolator_with_content_president_bush | 36.09 | ops/s |
| Mean Throughput | percolator_with_content_president_bush | 36.09 | ops/s |
| Median Throughput | percolator_with_content_president_bush | 36.09 | ops/s |
| Max Throughput | percolator_with_content_president_bush | 36.09 | ops/s |
| 100th percentile latency | percolator_with_content_president_bush | 35.9822 | ms |
| 100th percentile service time | percolator_with_content_president_bush | 7.93048 | ms |
| error rate | percolator_with_content_president_bush | 0 | % |
[...]
| Min Throughput | percolator_with_content_ignore_me | 16.1 | ops/s |
| Mean Throughput | percolator_with_content_ignore_me | 16.1 | ops/s |
| Median Throughput | percolator_with_content_ignore_me | 16.1 | ops/s |
| Max Throughput | percolator_with_content_ignore_me | 16.1 | ops/s |
| 100th percentile latency | percolator_with_content_ignore_me | 131.798 | ms |
| 100th percentile service time | percolator_with_content_ignore_me | 69.5237 | ms |
| error rate | percolator_with_content_ignore_me | 0 | % |
| Min Throughput | percolator_no_score_with_content_ignore_me | 29.37 | ops/s |
| Mean Throughput | percolator_no_score_with_content_ignore_me | 29.37 | ops/s |
| Median Throughput | percolator_no_score_with_content_ignore_me | 29.37 | ops/s |
| Max Throughput | percolator_no_score_with_content_ignore_me | 29.37 | ops/s |
| 100th percentile latency | percolator_no_score_with_content_ignore_me | 45.5703 | ms |
| 100th percentile service time | percolator_no_score_with_content_ignore_me | 11.316 | ms |
| error rate | percolator_no_score_with_content_ignore_me | 0 | % |
--------------------------------
[INFO] SUCCESS (took 18 seconds)
--------------------------------
Each task run by the percolator workload represents a specific OpenSearch API operation—such as Bulk or Search—that was performed when the test was run. Each task in the output summary contains the following information:
- Throughput: The number of successful OpenSearch operations per second.
- Latency: The amount of time, including wait time, taken for the request and the response to be sent and received by Benchmark.
- Service Time: The amount of time, excluding wait time, taken for the request and the response to be sent and received by Benchmark.
- Error Rate: The percentage of operations run during the task that were not successful or returned a 200 error code.
For more details about how the summary report is generated, see Summary report.
Running OpenSearch Benchmark on your own cluster
Now that you’re familiar with running OpenSearch Benchmark on a cluster, you can run OpenSearch Benchmark on your own cluster, using the same execute-test command, replacing the following settings.
- Replace
https://localhost:9200with your target cluster endpoint. This could be a URI likehttps://search.mydomain.comor aHOST:PORTspecification. - If the cluster is configured with basic authentication, replace the username and password in the command line with the appropriate credentials.
- Remove the
verify_certs:falsedirective if you are not specifyinglocalhostas your target cluster. This directive is needed only for clusters where SSL certificates are not set up. - If you are using a
HOST:PORTspecification and plan to use SSL/TLS, either specifyhttps://, or add theuse_ssl:truedirective to the--client-optionsstring option. - Remove the
--test-modeflag to run the full workload, rather than an abbreviated test.
You can copy the following command template to use in your own terminal:
opensearch-benchmark execute-test --pipeline=benchmark-only --workload=percolator --target-host=<OpenSearch Cluster Endpoint> --client-options=basic_auth_user:admin,basic_auth_password:admin
Next steps
See the following resources to learn more about OpenSearch Benchmark:
- User guide: Dive deep into how OpenSearch Benchmark can you help you track the performance of your cluster.
- Tutorials: Use step-by-step guides for more advanced Benchmarking configurations and functionality.