[BUG] Consul TTL heartbeat task can stop after one transient Consul error

### Current Behavior

Consul registry TTL heartbeats can stop permanently after one `agentCheckPass` failure.

`TtlScheduler.add()` schedules each heartbeat with `scheduleAtFixedRate(...)`:

```java
this.scheduler.scheduleAtFixedRate(new ConsulHeartbeatTask(instanceId), ttl, ttl, TimeUnit.SECONDS);
```

The task calls Consul directly and does not catch exceptions:

```java
TtlScheduler.this.client.agentCheckPass(this.checkId);
```

If Consul is temporarily unavailable or the request fails once, the exception escapes the fixed-rate task. `ScheduledThreadPoolExecutor` suppresses later executions of that periodic task after an uncaught exception.

### Expected Behavior

A transient Consul heartbeat failure should be logged and retried on the next TTL interval. Exceptions should not escape the scheduled heartbeat task.

### Impact

A registered instance can stop sending TTL pass updates after a single transient Consul error, causing Consul to mark it unhealthy until the instance re-registers or restarts.

### Code Location

- `shenyu-registry/shenyu-registry-consul/src/main/java/org/apache/shenyu/registry/consul/TtlScheduler.java`
- `add()` schedules the heartbeat task.
- `ConsulHeartbeatTask.run()` calls `agentCheckPass(...)` without exception handling.

Start in shenyu-registry/shenyu-registry-consul/src/main/java/org/apache/shenyu/registry/consul/TtlScheduler.java, reading add() and ConsulHeartbeatTask.run(). Check how scheduled-task exceptions affect later executions, then verify that a failed agentCheckPass call is logged and the heartbeat runs again on the next TTL interval without requiring re-registration or restart.

Written by the indexing model from the issue text.