lukaszraczylo/traefik-llmgateway
1
0
Fork 0
mirror of https://github.com/lukaszraczylo/traefik-llmgateway.git synced 2026-10-20 05:56:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(provider): list models in one page for Anthropic and Gemini

Browse Source 
Model discovery used each API's default page size: 20 for Anthropic and
50 for Gemini. A full catalog needed several sequential requests, and at
pod start that overran the 5s warm-fill budget. One replica then
logged "initial discovery ... context deadline exceeded" and went
without that provider's models until the next probe. Every page now
asks for the documented maximum of 1000: Anthropic limit and Gemini
pageSize. Gemini gets pageSize on each page, because it requires the
same parameters across one listing.

Claude-Session: https://claude.ai/code/session_01V5jYJNxpDFdJkyyExpADrv
This commit is contained in:
lukaszraczylo
2026-09-23 13:46:11 +01:00
parent 28722d983f
commit 8ac20d8d5e
4 changed files with 45 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
provider_anthropic_test.go
+9 -1
View File
@@ -471,9 +471,10 @@ func TestAnthropicAdapter_ListModels(t *testing.T) {
// page via ?after_id=<last_id> and return the union of both pages, not
// just the first page's 20 (Anthropic's own default page size).
func TestAnthropicAdapter_ListModels_FollowsPagination(t *testing.T) {
var gotAfterIDs []string
var gotAfterIDs, gotLimits []string
srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
gotAfterIDs = append(gotAfterIDs, r.URL.Query().Get("after_id"))
gotLimits = append(gotLimits, r.URL.Query().Get("limit"))
w.Header().Set("Content-Type", "application/json")
if r.URL.Query().Get("after_id") == "" {
_, _ = w.Write([]byte(`{"data":[{"id":"claude-opus-5"},{"id":"claude-sonnet-5"}],"has_more":true,"last_id":"claude-sonnet-5"}`))
@@ -504,6 +505,13 @@ func TestAnthropicAdapter_ListModels_FollowsPagination(t *testing.T) {
if len(gotAfterIDs) != 2 || gotAfterIDs[0] != "" || gotAfterIDs[1] != "claude-sonnet-5" {
t.Errorf("after_id sequence = %#v, want [\"\", \"claude-sonnet-5\"] (first page unqualified, second page's after_id = first page's last_id)", gotAfterIDs)
}
// Every page asks for Anthropic's maximum page size, so a real catalog
// arrives in one request inside warmFillTimeout.
for i, l := range gotLimits {
if l != "1000" {
t.Errorf("page %d limit = %q, want \"1000\"", i, l)
}
}
}
// TestAnthropicAdapter_ListModels_PageCapStopsRunawayPagination proves