Expose HTTP transient retry control and preserve shared capacity

This commit is contained in:
2026-09-26 15:31:31 +00:00
parent acf511fc0d
commit a4ef91fb7e
7 changed files with 167 additions and 10 deletions

View File

@@ -76,6 +76,10 @@ func (h *Handler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
writeError(w, http.StatusBadRequest, "invalid_request", "prompt_id is required")
return
}
if req.TransientRetries != nil && *req.TransientRetries < 0 {
writeError(w, http.StatusBadRequest, "invalid_request", "transient_retries must be greater than or equal to 0")
return
}
var mappedInputs map[string]promptkit.ArtifactRef
if len(req.Inputs) > 0 {
mappedInputs = make(map[string]promptkit.ArtifactRef, len(req.Inputs))
@@ -94,13 +98,14 @@ func (h *Handler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
}
res, err := h.runner.Run(r.Context(), promptkit.RunRequest{
PromptID: req.PromptID,
PromptVersion: req.PromptVersion,
ProfileID: req.ProfileID,
SessionID: req.SessionID,
Inputs: mappedInputs,
Vars: req.Vars,
Execution: model,
PromptID: req.PromptID,
PromptVersion: req.PromptVersion,
ProfileID: req.ProfileID,
SessionID: req.SessionID,
Inputs: mappedInputs,
Vars: req.Vars,
Execution: model,
TransientRetries: req.TransientRetries,
})
if err != nil {
status, code, message := mapRunError(err)
@@ -224,7 +229,7 @@ func mapRunError(err error) (int, string, string) {
return http.StatusBadRequest, "prompt_render_failed", "failed to render prompt"
case errors.Is(err, promptkit.ErrCapacityExceeded):
return http.StatusServiceUnavailable, "capacity_exceeded", "model backend capacity is exhausted"
case errors.Is(err, promptkit.ErrLLMGenerate):
case errors.Is(err, promptkit.ErrLLMGenerate), errors.Is(err, promptkit.ErrTransport), errors.Is(err, promptkit.ErrMalformedResponse), errors.Is(err, promptkit.ErrRetriesExhausted):
return http.StatusBadGateway, "llm_failed", "model generation request failed"
case errors.Is(err, promptkit.ErrValidation):
return http.StatusInternalServerError, "validation_runtime_failed", "validation runtime failed"