Auto-cleanup failed graphile jobs instead of keeping them around

Author: ericallam

apps/webapp/app/platform/zodWorker.server.ts

@@ -577,6 +577,17 @@ export class ZodWorker<TMessageCatalog extends MessageCatalogSchema> {
             span.recordException(new Error(String(error)));
           }
 
+          if (job.attempts >= job.max_attempts) {
+            logger.error("Job failed after max attempts", {
+              job,
+              attempts: job.attempts,
+              max_attempts: job.max_attempts,
+              error: error instanceof Error ? error.message : error,
+            });
+
+            return;
+          }
+
           throw error;
         } finally {
           span.end();

apps/webapp/app/services/apiRateLimit.server.ts

@@ -191,6 +191,7 @@ export const apiRateLimiter = authorizationRateLimitMiddleware({
   pathMatchers: [/^\/api/],
   // Allow /api/v1/tasks/:id/callback/:secret
   pathWhiteList: [
+    "/api/internal/stripe_webhooks",
     "/api/v1/authorization-code",
     "/api/v1/token",
     /^\/api\/v1\/tasks\/[^\/]+\/callback\/[^\/]+$/, // /api/v1/tasks/$id/callback/$secret

apps/webapp/app/services/worker.server.ts

@@ -232,11 +232,6 @@ function getWorkerQueue() {
   return new ZodWorker({
     name: "workerQueue",
     prisma,
-    cleanup: {
-      frequencyExpression: "13,27,43 * * * *",
-      ttl: env.WORKER_CLEANUP_TTL_DAYS * 24 * 60 * 60 * 1000, // X days
-      maxCount: 1000,
-    },
     runnerOptions: {
       connectionString: env.DATABASE_URL,
       concurrency: env.WORKER_CONCURRENCY,