feat(protocol): Add an error type#134
Merged
Merged
Conversation
The RoutedLlmClient implementations now have a concrete error `enum` to
return. This allows the caller to identify what happened. No more
`downcast`. Particularly, it allows the caller to identify timeouts
separately from other errors.
- Before RoutedLlmClient::call returned `Result<Response, Box<dyn Error + Send + Sync>>`
- Now it returns `Result<Response, RoutedLlmClientError>`
Which has these variants (see code for messages, this is edited):
```
pub enum RoutedLlmClientError {
InvalidRequest { message: String },
Configuration { message: String },
Transport { #[source] source: BoxError},
Timeout { #[source] source: BoxError },
ContextWindowExceeded { .. },
UpstreamHttp { status: u16, body: String, },
InvalidResponse { #[source] source: BoxError },
Other(#[from] BoxError),
}
```
Signed-off-by: Graham King <grahamk@nvidia.com>
Assisted-by: Codex:GPT 5.6 Sol medium
Contributor
Author
|
@coderabbitai full review |
✅ Action performedFull review finished. |
Thanks Claude! Signed-off-by: Graham King <grahamk@nvidia.com>
grahamking
marked this pull request as ready for review
July 24, 2026 15:43
WalkthroughThe PR replaces boxed routed-client errors with a typed ChangesTyped Routed Client Error Flow
Estimated code review effort: 4 (Complex) | ~45 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
crates/switchyard-server/src/lib.rs (1)
366-409: 🔒 Security & Privacy | 🟡 Minor | ⚡ Quick winFallback branch leaks the internal target name; use
source.to_string()for consistency.Every explicit arm here renders just the routed error's own message. The catch-all (
RoutedLlmClientError::Otherand any future unmatched variant) instead renderserror.to_string()on the outerLibsyError, whose#[error("client call to target {target:?} failed: {source}")]format embeds the internal target/model alias into the HTTP response body — inconsistent with the rest of the mapping and an avoidable bit of internal detail exposed to API callers.🔒 Suggested fix
- _ => server_error(error.to_string()), + _ => server_error(source.to_string()),🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/switchyard-server/src/lib.rs` around lines 366 - 409, Update the catch-all branch in algorithm_error to return server_error(source.to_string()) instead of error.to_string(), matching the other RoutedLlmClientError arms and avoiding exposure of the outer LibsyError target name.
🧹 Nitpick comments (1)
crates/libsy-llm-client/src/client.rs (1)
217-224: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick winDistinguish decode failures from body-read errors
Response::json()can fail on malformed JSON (is_decode()) or a body I/O problem (is_body()), but both currently fall through toResponseTranslation("invalid upstream JSON: ..."). Split those cases so transport failures stay inTransportand only actual JSON parse errors land inResponseTranslation.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@crates/libsy-llm-client/src/client.rs` around lines 217 - 224, Update the `http_response.json::<Value>()` error mapping in the response-handling branch to distinguish body I/O failures from JSON decode failures: preserve timeout mapping, map `error.is_body()` failures to `LlmClientError::Transport`, and map only `error.is_decode()` failures to `LlmClientError::ResponseTranslation`. Keep the existing invalid-JSON context for decode errors and handle any remaining error variant consistently with the client’s existing error model.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@crates/switchyard-server/src/lib.rs`:
- Around line 366-409: Update the catch-all branch in algorithm_error to return
server_error(source.to_string()) instead of error.to_string(), matching the
other RoutedLlmClientError arms and avoiding exposure of the outer LibsyError
target name.
---
Nitpick comments:
In `@crates/libsy-llm-client/src/client.rs`:
- Around line 217-224: Update the `http_response.json::<Value>()` error mapping
in the response-handling branch to distinguish body I/O failures from JSON
decode failures: preserve timeout mapping, map `error.is_body()` failures to
`LlmClientError::Transport`, and map only `error.is_decode()` failures to
`LlmClientError::ResponseTranslation`. Keep the existing invalid-JSON context
for decode errors and handle any remaining error variant consistently with the
client’s existing error model.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Enterprise
Run ID: 20cc504e-468d-44cc-a241-56001748c9af
⛔ Files ignored due to path filters (1)
Cargo.lockis excluded by!**/*.lock,!Cargo.lock
📒 Files selected for processing (17)
crates/libsy-llm-client/README.mdcrates/libsy-llm-client/src/client.rscrates/libsy-llm-client/src/error.rscrates/libsy/README.mdcrates/libsy/examples/ensemble.rscrates/libsy/examples/research_agent.rscrates/libsy/src/algorithms/fall_through.rscrates/libsy/src/algorithms/llm_class.rscrates/libsy/src/algorithms/rand.rscrates/libsy/src/algorithms/subagent_override.rscrates/libsy/src/core/algorithm.rscrates/libsy/src/error.rscrates/libsy/tests/observability.rscrates/protocol/Cargo.tomlcrates/protocol/src/client.rscrates/switchyard-py/src/libsy_bindings.rscrates/switchyard-server/src/lib.rs
Signed-off-by: Graham King <grahamk@nvidia.com>
grahamking
changed the title
messiaen
approved these changes
Jul 24, 2026
Signed-off-by: Graham King <grahamk@nvidia.com>
Signed-off-by: Graham King <grahamk@nvidia.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The RoutedLlmClient implementations now have a concrete error
enumtoreturn. This allows the caller to identify what happened. No more
downcast. Particularly, it allows the caller to identify timeoutsseparately from other errors.
Result<Response, Box<dyn Error + Send + Sync>>Result<Response, LlmClientError>Which has these variants (see code for messages, this is edited) (note some more added in later commit):
libsy-llm-clientuses this error type directly.Signed-off-by: Graham King grahamk@nvidia.com
Assisted-by: Codex:GPT 5.6 Sol medium