tidb should return more correct info after down pd leader

## Bug Report

Please answer these questions before submitting your issue. Thanks!

### 1. Minimal reproduce step (Required)
restart pd leader pod.

### 2. What did you expect to see? (Required)
query failed with message that failed to connect to pd.

### 3. What did you see instead (Required)
[2022-12-21 07:31:30] [2022-12-21 07:31:30] execute run failed, err exec UPDATE order_line SET ol_delivery_d = ? WHERE (ol_w_id, ol_d_id, ol_o_id) IN (
[2022-12-21 07:31:30] (?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?),(?,?,?)
[2022-12-21 07:31:30] ) failed Error 1105: rpc error: code = Unknown desc = server not started
[2022-12-21 07:31:30] [2022-12-21 07:31:30] execute run failed, err exec SELECT no_o_id FROM new_order WHERE no_w_id = ? AND no_d_id = ? ORDER BY no_o_id ASC LIMIT 1 FOR UPDATE failed Error 1105: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 172.30.168.13:2379: connect: connection refused"

### 4. What is your TiDB version? (Required)
6.5