Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Amend the endpoint picker protocol to support fallbacks and subsetting #445

Merged
merged 9 commits into from
Mar 13, 2025
44 changes: 37 additions & 7 deletions docs/proposals/004-endpoint-picker-protocol/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,27 +9,57 @@ This doc defines the protocol between the EPP and the proxy (e.g, Envoy).
The EPP MUST implement the Envoy
[external processing service](https://www.envoyproxy.io/docs/envoy/latest/api-v3/service/ext_proc/v3/external_processor) protocol.

## Endpoint Subset
For each HTTP request, the proxy CAN communicate the subset of endpoints the EPP MUST pick from by setting an unstructured entry in the [filter metadata](https://github.com/envoyproxy/go-control-plane/blob/63a55395d7a39a8d43dcc7acc3d05e4cae7eb7a2/envoy/config/core/v3/base.pb.go#L819) field of the ext-proc request. The metadata entry for the subset list MUST be wrapped with an outer key (which represents the metadata namespace) with a default of `envoy.lb.subset_hint`.

```go
filterMetadata: {
"envoy.lb.subset_hint" {
"x-gateway-destination-endpoint-subset": [<ip:port>, <ip:port>, ...]
}
}
```

If the key `x-gateway-destination-endpoint-subset` is set, the EPP MUST only select endpoints from the specified list. If none of the endpoints in the list is eligible or the list is empty, then the EPP MUST return a 429 status code. If the EPP does not select from the list, then this leads to unpredictable behavior.

If the key `x-gateway-destination-endpoint-subset` is not set, then the EPP MUST select from the set defined by the `InferencePool` selector.

## Destination Endpoint
For each HTTP request, the EPP MUST communicate to the proxy the picked model server endpoint via:

1. Setting the `x-gateway-destination-endpoint` HTTP header to the selected endpoint in <ip:port> format.

2. Set an unstructured entry in the [dynamic_metadata](https://github.com/envoyproxy/go-control-plane/blob/c19bf63a811c90bf9e02f8e0dc1dcef94931ebb4/envoy/service/ext_proc/v3/external_processor.pb.go#L320) field of the ext-proc response. The metadata entry for the picked endpoint MUST be wrapped with an outer key (which represents the metadata namespace) with a default of `envoy.lb`.

The final metadata necessary would look like:
The primary endpoint MUST be set using the key `x-gateway-destination-endpoint` as follows:
```go
dynamicMetadata: {
"envoy.lb": {
"x-gateway-destination-endpoint": <ip:port>"
"x-gateway-destination-endpoint": <ip:port>
}
}
```

Note:
- If the EPP did not communicate the server endpoint via these two methods, it MUST return an error.
Constraints:
- If the EPP did not communicate the server endpoint via these two methods, it MUST return an error as follows:
- 503 (Serivce Unavailable) if there are no ready endpoints.
- 429 (Too Many Requests) if the request should be dropped (e.g., a Sheddable request, and the servers under heavy load).
- The EPP MUST not set two different values in the header and the inner response metadata value.
- Setting different value leads to unpredictable behavior because proxies aren't guaranteed to support both paths, and so this protocol does not define what takes precedence.

### Destination endpoint fallback
A single fallback endpoint CAN be set using the key `x-gateway-destination-endpoint-fallback` in the same metadata namespace as one used for `x-gateway-destination-endpoint` as follows:

## Why envoy.lb namespace as a default?
The `envoy.lb` namesapce is a predefined namespace used for subsetting. One common way to use the selected endpoint returned from the server, is [envoy subsets](https://www.envoyproxy.io/docs/envoy/latest/intro/arch_overview/upstream/load_balancing/subsets) where host metadata for subset load balancing must be placed under `envoy.lb`.
```go
dynamicMetadata: {
"envoy.lb" {
"x-gateway-destination-endpoint-fallback": <ip:port>
}
}
```

Setting different value leads to unpredictable behavior because proxies aren't guaranteed to support both paths, and so this protocol does not define what takes precedence.
### Why envoy.lb namespace as a default?
The `envoy.lb` namespace is a predefined namespace. One common way to use the selected endpoint returned from the server, is [envoy subsets](https://www.envoyproxy.io/docs/envoy/latest/intro/arch_overview/upstream/load_balancing/subsets) where host metadata for subset load balancing must be placed under `envoy.lb`. Note that this is not related to the subsetting feature discussed above, this is an enovy implementation detail.

## Matching An InferenceModel
The model name of a request MUST match the `Sepc.ModelName` parameter of one of the `InferenceModels` referencing the `InferencePool` managed by the EPP. Otherwise, the EPP MUST return a 404 status code.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

typo: "Sepc"

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done