feat(webapp): per-client database pool and connect timeout overrides

Adds optional per-client env overrides for the Prisma pool_timeout and connect_timeout, one pair each for the writer and read replica of both the primary and run-ops databases, falling back to the shared DATABASE_POOL_TIMEOUT / DATABASE_CONNECTION_TIMEOUT when unset.

This lets a fail-fast connect timeout be set for one backend's clients without changing the others. No behavior change until an override is set.

Author: ericallam

.server-changes/per-client-connection-timeouts.md

@@ -0,0 +1,6 @@
+---
+area: webapp
+type: improvement
+---
+
+Allow the database connection pool and connect timeouts to be tuned separately for each writer and read replica, falling back to the shared defaults when unset.

apps/webapp/app/db.server.ts

@@ -398,8 +398,10 @@ export function buildWriterClient({
 }): PrismaClient {
   const databaseUrl = buildPrismaConnectionUrl(url, {
     connectionLimit: env.DATABASE_CONNECTION_LIMIT.toString(),
-    poolTimeout: env.DATABASE_POOL_TIMEOUT.toString(),
-    connectTimeout: env.DATABASE_CONNECTION_TIMEOUT.toString(),
+    poolTimeout: (env.DATABASE_WRITER_POOL_TIMEOUT ?? env.DATABASE_POOL_TIMEOUT).toString(),
+    connectTimeout: (
+      env.DATABASE_WRITER_CONNECTION_TIMEOUT ?? env.DATABASE_CONNECTION_TIMEOUT
+    ).toString(),
     applicationName: env.SERVICE_NAME,
   });
 
@@ -545,8 +547,10 @@ export function buildReplicaClient({
 }): PrismaClient {
   const replicaUrl = buildPrismaConnectionUrl(url, {
     connectionLimit: env.DATABASE_CONNECTION_LIMIT.toString(),
-    poolTimeout: env.DATABASE_POOL_TIMEOUT.toString(),
-    connectTimeout: env.DATABASE_CONNECTION_TIMEOUT.toString(),
+    poolTimeout: (env.DATABASE_READ_REPLICA_POOL_TIMEOUT ?? env.DATABASE_POOL_TIMEOUT).toString(),
+    connectTimeout: (
+      env.DATABASE_READ_REPLICA_CONNECTION_TIMEOUT ?? env.DATABASE_CONNECTION_TIMEOUT
+    ).toString(),
     applicationName: env.SERVICE_NAME,
   });
 
@@ -675,8 +679,10 @@ function buildRunOpsWriterClient({
 }): RunOpsPrismaClient {
   const databaseUrl = buildPrismaConnectionUrl(url, {
     connectionLimit: env.DATABASE_CONNECTION_LIMIT.toString(),
-    poolTimeout: env.DATABASE_POOL_TIMEOUT.toString(),
-    connectTimeout: env.DATABASE_CONNECTION_TIMEOUT.toString(),
+    poolTimeout: (env.RUN_OPS_DATABASE_WRITER_POOL_TIMEOUT ?? env.DATABASE_POOL_TIMEOUT).toString(),
+    connectTimeout: (
+      env.RUN_OPS_DATABASE_WRITER_CONNECTION_TIMEOUT ?? env.DATABASE_CONNECTION_TIMEOUT
+    ).toString(),
     applicationName: env.SERVICE_NAME,
   });
 
@@ -728,8 +734,12 @@ function buildRunOpsReplicaClient({
     connectionLimit: (
       env.RUN_OPS_DATABASE_READ_REPLICA_CONNECTION_LIMIT ?? env.DATABASE_CONNECTION_LIMIT
     ).toString(),
-    poolTimeout: env.DATABASE_POOL_TIMEOUT.toString(),
-    connectTimeout: env.DATABASE_CONNECTION_TIMEOUT.toString(),
+    poolTimeout: (
+      env.RUN_OPS_DATABASE_READ_REPLICA_POOL_TIMEOUT ?? env.DATABASE_POOL_TIMEOUT
+    ).toString(),
+    connectTimeout: (
+      env.RUN_OPS_DATABASE_READ_REPLICA_CONNECTION_TIMEOUT ?? env.DATABASE_CONNECTION_TIMEOUT
+    ).toString(),
     applicationName: env.SERVICE_NAME,
   });
 

apps/webapp/app/env.server.ts

@@ -120,6 +120,10 @@ const EnvironmentSchema = z
     DATABASE_CONNECTION_LIMIT: z.coerce.number().int().default(10),
     DATABASE_POOL_TIMEOUT: z.coerce.number().int().default(60),
     DATABASE_CONNECTION_TIMEOUT: z.coerce.number().int().default(20),
+    DATABASE_WRITER_POOL_TIMEOUT: z.coerce.number().int().optional(),
+    DATABASE_WRITER_CONNECTION_TIMEOUT: z.coerce.number().int().optional(),
+    DATABASE_READ_REPLICA_POOL_TIMEOUT: z.coerce.number().int().optional(),
+    DATABASE_READ_REPLICA_CONNECTION_TIMEOUT: z.coerce.number().int().optional(),
     // Dashboard-agent conversation store. Cloud points this at a dedicated
     // database; when unset it falls back to DATABASE_URL (OSS), where
     // the tables live in the isolated `trigger_dashboard_agent` schema.
@@ -186,6 +190,10 @@ const EnvironmentSchema = z
       .optional(),
     // Optional cap for the unpooled new run-ops read replica. Unset falls back to DATABASE_CONNECTION_LIMIT.
     RUN_OPS_DATABASE_READ_REPLICA_CONNECTION_LIMIT: z.coerce.number().int().optional(),
+    RUN_OPS_DATABASE_WRITER_POOL_TIMEOUT: z.coerce.number().int().optional(),
+    RUN_OPS_DATABASE_WRITER_CONNECTION_TIMEOUT: z.coerce.number().int().optional(),
+    RUN_OPS_DATABASE_READ_REPLICA_POOL_TIMEOUT: z.coerce.number().int().optional(),
+    RUN_OPS_DATABASE_READ_REPLICA_CONNECTION_TIMEOUT: z.coerce.number().int().optional(),
     // Direct DSN for applying the full @trigger.dev/database migrations to the LEGACY run-ops DB, keeping
     // its schema current after the control plane moves off it. Direct, not pooled — migrations never run
     // over a pooler. Optional; unset -> the entrypoint's legacy migrate step is skipped.