feat: Phase 2 - reliability & observability
CI / Lint (push) Has been cancelled
CI / Test (push) Has been cancelled
CI / Build (push) Has been cancelled

- Circuit breaker: proper thresholds (3 failures, 30s timeout)
- HTTP timeouts: 30s on all providers (was no timeout)
- Structured logging: slog replaces fmt.Printf throughout
- Stream errors: propagated as SSE error events to client
- Registry fetch: retry with backoff (3 attempts)
- Registry reads in dashboard protected by RWMutex
This commit is contained in:
2026-04-26 14:48:41 -04:00
parent 8a8d8d1477
commit 1f574d8134
15 changed files with 126 additions and 608 deletions
+28 -21
View File
@@ -6,38 +6,48 @@ import (
"log"
"time"
"gophergate/internal/models"
"github.com/go-resty/resty/v2"
"gophergate/internal/models"
)
const ModelsDevURL = "https://models.dev/api.json"
func FetchRegistry() (*models.ModelRegistry, error) {
log.Printf("Fetching model registry from %s", ModelsDevURL)
client := resty.New().SetTimeout(10 * time.Second)
resp, err := client.R().Get(ModelsDevURL)
if err != nil {
return nil, fmt.Errorf("failed to fetch registry: %w", err)
var lastErr error
for attempt := 0; attempt < 3; attempt++ {
if attempt > 0 {
backoff := time.Duration(1<<attempt) * time.Second
time.Sleep(backoff)
}
resp, err := client.R().Get(ModelsDevURL)
if err != nil {
lastErr = fmt.Errorf("attempt %d: %w", attempt+1, err)
continue
}
if !resp.IsSuccess() {
lastErr = fmt.Errorf("attempt %d: HTTP %d", attempt+1, resp.StatusCode())
continue
}
var providers map[string]models.ProviderInfo
if err := json.Unmarshal(resp.Body(), &providers); err != nil {
lastErr = fmt.Errorf("attempt %d: unmarshal: %w", attempt+1, err)
continue
}
log.Println("Successfully loaded model registry")
return &models.ModelRegistry{Providers: providers}, nil
}
if !resp.IsSuccess() {
return nil, fmt.Errorf("failed to fetch registry: HTTP %d", resp.StatusCode())
}
var providers map[string]models.ProviderInfo
if err := json.Unmarshal(resp.Body(), &providers); err != nil {
return nil, fmt.Errorf("failed to unmarshal registry: %w", err)
}
log.Println("Successfully loaded model registry")
return &models.ModelRegistry{Providers: providers}, nil
return nil, fmt.Errorf("failed to fetch registry after 3 attempts: %w", lastErr)
}
func CalculateCost(registry *models.ModelRegistry, modelID string, promptTokens, completionTokens, reasoningTokens, cacheRead, cacheWrite uint32) float64 {
meta := registry.FindModel(modelID)
if meta == nil || meta.Cost == nil {
log.Printf("[DEBUG] CalculateCost: model %s not found or has no cost metadata", modelID)
return 0.0
}
@@ -62,8 +72,5 @@ func CalculateCost(registry *models.ModelRegistry, modelID string, promptTokens,
cost += float64(cacheWrite) * (*meta.Cost.CacheWrite) / 1000000.0
}
log.Printf("[DEBUG] CalculateCost: model=%s, uncached=%d, completion=%d, reasoning=%d, cache_read=%d, cache_write=%d, cost=%f (input_rate=%f, output_rate=%f)",
modelID, uncachedTokens, completionTokens, reasoningTokens, cacheRead, cacheWrite, cost, meta.Cost.Input, meta.Cost.Output)
return cost
}