Update performance benchmarks #3818
Conversation
Signed-off-by: yaron2 <[email protected]>
| @@ -19,7 +19,7 @@ For applications using actors in Dapr there are two aspects to be considered. Fi | |||
| * Sidecar Injector (control plane) | |||
| * Sentry (optional, control plane) | |||
|
|
|||
| ## Performance summary for Dapr v1.0 | |||
| ## Performance summary for Dapr v1.12 | |||
|
|
|||
| The actors API in Dapr sidecar will identify which hosts are registered for a given actor type and route the request to the appropriate host for a given actor ID. The host runs an instance of the application and uses the Dapr SDK (.Net, Java, Python or PHP) to handle actors requests via HTTP. | |||
There was a problem hiding this comment.
| The actors API in Dapr sidecar will identify which hosts are registered for a given actor type and route the request to the appropriate host for a given actor ID. The host runs an instance of the application and uses the Dapr SDK (.Net, Java, Python or PHP) to handle actors requests via HTTP. | |
| The actors API in Dapr sidecar identifies which hosts are registered for a given actor type and routes the request to the appropriate host for a given actor ID. The host runs an instance of the application and uses the Dapr SDK (.Net, Java, Python, Go) to handle actors requests via HTTP. |
There was a problem hiding this comment.
Is there a link to the Actor performance test that we can include, so that people can try this themselves?
|
|
||
| ### Kubernetes performance test setup | ||
|
|
||
| The test was conducted on a 3 node Kubernetes cluster, using commodity hardware running 4 cores and 8GB of RAM, without any network acceleration. |
There was a problem hiding this comment.
Is there a link to the performance test that we can include, so that people can try this themselves?
|
|
||
| <img src="/images/perf_invocation_p99.png" alt="Latency for 99th percentile"> | ||
| This performance is on par or better than commonly used service meshes. |
There was a problem hiding this comment.
Let's quantify this with a link to service meshes perf results, otherwise this is hearsay and we should not include this
There was a problem hiding this comment.
In removing the graphs, there is not mTLS measurement comparison. What impact does mTLS have on the performance tests, it would be good to at least clarify this with a % increase. The majority of people will use mTLS and so this number is important.
| <img src="/images/perf_invocation_p90.png" alt="Latency for 90th percentile"> | ||
|
|
||
| <br> | ||
| The Dapr sidecar uses 0.45 vCPU and 38Mb per 1000 requests per second. |
There was a problem hiding this comment.
| The Dapr sidecar uses 0.45 vCPU and 38Mb per 1000 requests per second. | |
| The Dapr sidecar uses 0.45 vCPU and 38Mb per 1000 requests per second. | |
| | Sentry | 0.005 | 13.6 Mb | ||
| | Sidecar Injector | 0.002 | 14.6 Mb | ||
| | Placement | 0.001 | 20.9 Mb | ||
| | Operator | 0.003 | 18 Mb |
There was a problem hiding this comment.
Is this the only place that we capture the Control Plane numbers and are these consistently the same for all the performance test? If so we should separate these out. If not, then should each perf test have different control plane numbers like this one?
| type: docs | ||
| title: "Pub/sub performance" | ||
| linkTitle: "Pub/sub performance" | ||
| weight: 20000 |
There was a problem hiding this comment.
| weight: 20000 | |
| weight: 30000 |
| type: docs | ||
| title: "State performance" | ||
| linkTitle: "State performance" | ||
| weight: 20000 |
There was a problem hiding this comment.
| weight: 20000 | |
| weight: 40000 |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
@yaron2 friendly bump |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
|
Stale PR, paging all reviewers |
Closes #3468