grpc.io/content/en/blog/performance-benchmarks-gke.md

360 lines
17 KiB
Markdown

---
title: gRPC performance benchmarks on GKE
date: 2022-03-01
authors:
- name: Paulo Castello da Costa
link: https://github.com/paulosjca
- name: Jan Tattermusch
link: https://github.com/jtattermusch
---
[gRPC performance benchmarks][benchmarking] have now been transitioned to run on
GKE, with similar results but much increased flexibility.
## Background
gRPC performance testing requires a test driver and workers (one or more clients
and a server), as described in [gRPC performance benchmarks][benchmarking]. Each
test may have a different configuration, or _scenario_, that is passed to the
driver and specified as a JSON file. Previously, the driver was run by the
continuous integration process, and the workers were run on long-lived GCE VMs.
This gave rise to several limitations:
1. Tests ran sequentially and were difficult to parallelize, since they ran on
(the same) fixed VMs.
1. The state of the VMs was not guaranteed to be the same at the start of each
test.
1. Running manual experiments required configuring new VMs, which was a manual
process, or reusing existing VMs, with the risks of collision with other
users and of having VMs in an unknown state.
## Benchmarking on Kubernetes
The core of the current framework is a [custom controller][] for managing
Kubernetes resources of kind [LoadTest][loadtest]. This controller must be
deployed to a Kubernetes cluster before load tests can be run on it. The
controller is implemented with [kubebuilder][]. The code for the controller is
stored in the [Test Infra][testinfra] repository. For further documentation on
individual LoadTest fields, see the [LoadTest
implementation][loadtestimplementation].
LoadTest configurations specify driver, client and server pods to be created for
the test. Once the configuration is applied to the cluster (for instance, with
`kubectl apply -f`), the controller will create the pods and the test will run.
If multiple configurations are applied to the cluster, the controller will
create pods as long as there are resources available, allowing tests to run in
parallel.
[Examples][examples] include basic configurations that can be applied directly,
and templates that require additional steps and parameter substitution.
- Basic configurations rely on **clone**, **build** and **runtime** worker
images that are bundled with each release of the controller. The clone and
build images are used to build gRPC binaries that are passed to the runtime
container. These configurations are suitable as examples and for one-off
testing.
- Template configurations rely on worker images that are built before starting
the tests. These **prebuilt images** include the gRPC binary, eliminating the
need to clone and build before each test. Template substitution is used to
point to the location of the worker images. These configurations are suitable
for running a batch of tests on the same gRPC version, or for running the same
test repeatedly.
In addition to the controller, the [Test Infra][testinfra] repository contains a
set of [tools][], including a test runner and tools to build and delete prebuilt
worker images, as well as a [dashboard](#dashboard) implementation.
The tools related to prebuilt workers use `gcloud` internally and are dependent
on GKE. Other than that, all components of the framework are built on Kubernetes
itself and independent of GKE. That is, it should be possible to deploy the
controller and run tests on a custom Kubernetes cluster or on another cloud
provider's Kubernetes offering.
[custom controller]:
https://github.com/grpc/test-infra/blob/master/cmd/controller/main.go
[kubebuilder]: https://kubebuilder.io
[loadtest]:
https://github.com/grpc/test-infra/blob/master/config/crd/bases/e2etest.grpc.io_loadtests.yaml
[loadtestimplementation]:
https://github.com/grpc/test-infra/blob/master/api/v1/loadtest_types.go
## Cluster setup
The cluster running benchmark jobs must be configured with node pools
dimensioned for the number of simultaneous tests that it should support. The
controller uses `pool` as a node selector for the various pod types. Worker pods
have mutual anti-affinity, so one node is required per pod.
For example, the node pools that are used in our continuous integration setup
are configured as follows:
| Pool name | Node count | Machine type | Kubernetes labels |
| :------------------- | ---------: | :------------- | :----------------------------------------- |
| system | 2 | e2-standard-8 | default-system-pool:true, pool:system |
| drivers-ci | 8 | e2-standard-2 | pool:drivers-ci |
| workers-c2-8core-ci | 8 | c2-standard-8 | pool:workers-c2-8core-ci |
| workers-c2-30core-ci | 8 | c2-standard-30 | pool:workers-c2-30core-ci |
Since each scenario in our tests requires one driver and two workers, this
configuration supports four simultaneous tests on 8-core machines and four on
30-core machines. Drivers require few resources, and do not have mutual
anti-affinity. We find it convenient to schedule them on two-core machines with
a node count set to the required number of drivers, rather than on a larger
shared machine, because that allows the driver pool to be resized together with
the worker pools. The controller itself is scheduled in the `system` pool.
In addition to the pools used in continuous integration, our cluster contains
additional node pools that can be used for ad hoc testing:
| Pool name | Node count | Machine type | Kubernetes labels |
| :------------- | ---------: | :------------- | :------------------------------------------------ |
| drivers | 8 | e2-standard-8 | default-driver-pool:true, pool:drivers |
| workers-8core | 8 | e2-standard-8 | default-worker-pool:true, pool:workers-8core |
| workers-32core | 8 | e2-standard-32 | pool:workers-32core |
Some pools are labeled with `default-*-pool` labels. These labels specify which
pool to use if it is not specified in the LoadTest configuration. With the
configuration above, these tests (for instance, the tests specified in the
[examples][]) will use the `drivers` and `workers-8core` pools, and not
interfere with continuous integration jobs. The default labels are defined as
part of the controller build: if they are not set, the controller will only run
tests where the `pool` labels are specified explicitly.
## Controller deployment
The steps for building and deploying a controller are described in the
[deployment documentation][].
[deployment documentation]:
https://github.com/grpc/test-infra/blob/master/doc/deployment.md
## Continuous integration
Our continuous integration setup is described in the [gRPC OSS benchmarks
README][] in the [gRPC Core] repository. The main continuous integration job
uses the script [grpc_e2e_performance_gke.sh][] to generate the data presented
on the dashboard linked to the [gRPC performance benchmarks][benchmarking] page.
Each continuous integration run has three phases:
1. Generate test configurations.
1. Build and push worker images.
1. Run the tests.
Each continuous integration run executes 122 tests using the 8-core worker pool,
and 98 test using the 30-core worker pool. Each test runs one test scenario.
Tests using C++, C#, Java and Python workers run on both pools. Tests using
Node.js, PHP and Ruby workers run only on the 8-core pool.
[Config generation](#config-generation) for all these combinations takes
negligible time (~1s).
The configurations used in continuous integration require worker images that
include the gRPC binaries to be tested. These images depend only on the worker's
language, so these [prebuilt images](#prebuilt-images) are built in advance and
pushed to an image repository. This process takes approximately 20 minutes.
A [test runner](#test-runner) manages the rate at which tests are applied to the
cluster, collects test results and logs, and deletes tests once they complete
successfully. Two tests at a time are allowed to run on each pool. This phase
takes approximately 50 minutes to complete.
Each test scenario is configured to take 30s to run, plus a 5s warm-up period
(15s for Java). This places a lower bound on the time needed to run each test.
The observed run time of the 122 tests in the 8-core pool (16 of them in Java),
running two tests at a time, implies that the overhead introduced by pod
creation and deletion is modest, at ~12.8s per test.
[grpc_e2e_performance_gke.sh]:
https://github.com/grpc/grpc/blob/master/tools/internal_ci/linux/grpc_e2e_performance_gke.sh
### Config generation
Since we are running hundreds of tests that mostly share the same components
(driver and workers for the various languages), it becomes necessary to generate
configurations that includes repetitive driver and worker configurations and
differ only in the scenario being tested. In addition, each configuration must
have a unique name, since that is a requirement for resources applied to a
Kubernetes cluster.
We handle these issues by [generating load test configurations][] with a tool.
The tool is stored in the [gRPC Core] repository, where test scenarios are also
defined.
[generating load test configurations]:
https://github.com/grpc/grpc/blob/master/tools/run_tests/performance/README.md#generating-load-test-configurations
### Prebuilt images
The configurations generated for continuous integration use a set of prebuilt
images. These images are built and pushed to an image repository before running
tests. The images are deleted at the end of each test run.
For details on the tools used to prepare and delete images, see [Using prebuilt
images with gRPC OSS benchmarks][prebuilt].
[prebuilt]:
https://github.com/grpc/test-infra/blob/master/tools/README.md#using-prebuilt-images-with-grpc-oss-benchmarks
### Test runner
The [test runner][] takes the test configurations generated previously, applies
each configuration to the cluster, polls each LoadTest resource for completion,
collects artifacts such as results and pod logs, and (optionally) deletes
resources once each test completes successfully.
The test runner maintains separate _queues_ for tests that require the same
resources on the cluster (e.g. 8-core or 30-core worker nodes). Tests
configurations belonging to the same queue are not applied ot the cluster at
once, but according to a _concurrency level_ set for each queue. Our continuous
integration tests run in two queues (corresponding to 8-core and 30-core worker
nodes). The concurrency level of each queue is set to two.
Once a configuration is applied to the cluster, the controller creates client,
driver and server pods to run the test, monitors test execution, and updates the
status of the LoadTest resource.
The design of the test runner can be explained as follows:
1. The use of the test runner allows the continuous integration job to wait for
all tests to complete, collect tests artifacts, and prepare a report with
results.
1. The use of separate queues (indicated by annotations in each test
configuration) allows tests that do not require the same cluster resources to
be managed independently of each other.
1. The use of limited concurrency levels reduces the number of tests applied to
the cluster at a time. This has several benefits:
1. The load on the test runner is reduced, since there are fewer LoadTest
resources on the cluster at a time, and the runner polls these resources
periodically for completion. The polling interval in our continuous
integration is set to 5s.
1. The load on the controller is reduced, since there are fewer LoadTest
resources at a time for it to control.
1. Each test can have a shorter timeout period, since the time takes for the
controller to start each test is more predictable. Timeouts are necessary
to account for error cases where client or server pods hang and prevent
the test from completing. These cases are rare but can accumulate and
consume cluster resources, preventing other tests from running. Tests in
our continuous integration have a timeout of 15 minutes.
1. The concurrency level can be set lower than the capacity of the cluster,
allowing a user to run a batch of tests without preventing other users
from running tests simultaneously.
1. The option of deleting each test once it completes successfully (and after
results and logs are collected) provides better control of the lifecycle of
each test.
1. The default behavior of the controller is to keep a LoadTest resource and
associated pods on the cluster until it reaches a set TTL, and then delete
it. Our continuous integration specifies a TTL of 24h for each test.
1. Pods belonging to a completed LoadTest are in terminated state, and so
consume no resources on the cluster. However, terminated pods can be
garbage-collected at any time.
1. If we let pods belonging to all completed tests stay in our continuous
integration cluster, we find that they are garbage-collected within one
hour.
1. If we delete LoadTest resources for tests that complete successfully, the
associated pods are also deleted. In this case, the pods belonging to
_unsuccessful_ tests, which are few in number and may be useful for
debugging, stay on the cluster until the 24h TTL is reached.
[test runner]:
https://github.com/grpc/test-infra/blob/master/tools/README.md#test-runner
### Dashboard
Test results from continuous integration are saved to [BigQuery][bigquery]. The
data stored in BigQuery is then replicated to a Postgres database for
visualization on a dashboard.
The code for the dashboard, as well as the configuration of the main continuous
integration dashboard, are stored in the [Test Infra][testinfra] repository.
This brings the following benefits:
1. The main dashboard is maintained by updating the stored configuration,
instead of updating it directly in the UI.
1. Users can deploy their own dashboard, using their own configuration.
This contrasts with the dashboard for the previous benchmarks, built using
Perfkit Explorer, which was maintained by updating it directly in the UI, and
could not be easily replicated by users.
For details, see [dashboard implementation][].
[![Dashboard snapshot][snapshot]<span class="hk-no-external-icon"></span>](https://grafana-dot-grpc-testing.appspot.com)
[bigquery]: https://cloud.google.com/bigquery
[dashboard implementation]:
https://github.com/grpc/test-infra/tree/master/dashboard/README.md
[snapshot]: /img/blog/performance-benchmarks-gke/dashboard.png
## Results
The following observations can be made on the results and user experience of the
gRPC benchmarking on GKE:
1. Performance metrics (latency, QPS, etc.) produce the same or better results
as the old benchmarks on GCE.
1. The overhead of pod creation and deletion for each test in GKE is small (less
than 15s) in our benchmarking cluster.
1. Test images are dockerized and started again for each test, resulting in
several benefits:
1. Results are more consistent.
1. Runtime errors are rare.
1. System is split into well-defined components, resulting in simpler
upgrades.
1. Tests can be easily parallelized, resulting in faster execution times.
1. Experiments are easier to perform.
Examples of best practices and insights derived from experimentation:
1. Use `c2` instances for clients and servers (instance types matter a lot for
the observed latency and its variance and also for the measured throughput).
1. GKE pod-to-pod networking only has a very small overhead over raw GCE
networking. You can get raw GCE networking performance by setting
`hostnetworking:true` for the benchmark pods.
1. For Java under Docker, the JVM may not be able to detect the number of
available processors automatically. This can lead to very pessimistic
results, since gRPC uses the detected number of processors for sizing thread
pools for processing events. A workaround is to set the number of processors
explicitly. This workaround is implemented
[here](https://github.com/grpc/test-infra/pull/231).
## Running your own
The code in the [Test Infra][testinfra] repository allows any user to create a
cluster, deploy a controller, run gRPC benchmarks, and display results on their
own dashboards. If you are interested in performance, and run your own
benchmarks, [let us know!][g/grpc-io]
[benchmarking]: /docs/guides/benchmarking
[examples]:
https://github.com/grpc/test-infra/blob/master/config/samples/README.md
[g/grpc-io]: https://groups.google.com/g/grpc-io
[grpc core]: https://github.com/grpc/grpc
[grpc oss benchmarks readme]:
https://github.com/grpc/grpc/blob/master/tools/run_tests/performance/README.md#grpc-oss-benchmarks
[testinfra]: https://github.com/grpc/test-infra
[tools]: https://github.com/grpc/test-infra/blob/master/tools/README.md