fix: restore 30-minute timeout for Bazel builds (#19609)

I think raising it to 45 minutes in
https://github.com/openai/codex/pull/19578 was a mistake for the reasons
explained in the comments in the code. Instead, we attempt to defend
against timeouts by increasing the number of shards in
`app-server-all-test` so that a "true failure" that gets run 3x should
not take as much wall clock time.

codex-rs/app-server/BUILD.bazel

@@ -5,7 +5,13 @@ codex_rust_crate(
     crate_name = "codex_app_server",
     integration_test_timeout = "long",
     test_shard_counts = {
-        "app-server-all-test": 8,
+        # Note app-server-all-test has a large number of integration tests, so
+        # even a single shard can be quite slow. When there is a legitimate
+        # test failure in a shard, it will still get run 3x in total, which
+        # can cause us to exhaust our CI timeout if the shard happens to run
+        # long. Using a higher shard count for app-server-all-test should help
+        # mitigate this risk.
+        "app-server-all-test": 16,
         "app-server-unit-tests": 8,
     },
     test_tags = ["no-sandbox"],