the time of job needed will be count for every ?

```python
def job():
global count
logger.info("I'm doing a periodic job!" + str(count))
count = count + 1
time.sleep(30)

# 每隔 2 分钟执行一次任务
job()
schedule.every(2).minutes.do(job)
while True:
# logger.info("waiting for")
schedule.run_pending()
time.sleep(1)
```

the log is

```shell
2024-03-18 12:03:44,170 - INFO - I'm doing a periodic job!0
2024-03-18 12:06:14,534 - DEBUG - Running job Job(interval=2, unit=minutes, do=job, args=(), kwargs={})
2024-03-18 12:06:14,535 - INFO - I'm doing a periodic job!1
2024-03-18 12:08:44,888 - DEBUG - Running job Job(interval=2, unit=minutes, do=job, args=(), kwargs={})
2024-03-18 12:08:44,889 - INFO - I'm doing a periodic job!2
2024-03-18 12:11:15,267 - DEBUG - Running job Job(interval=2, unit=minutes, do=job, args=(), kwargs={})
2024-03-18 12:11:15,267 - INFO - I'm doing a periodic job!3
2024-03-18 12:13:45,660 - DEBUG - Running job Job(interval=2, unit=minutes, do=job, args=(), kwargs={})
2024-03-18 12:13:45,660 - INFO - I'm doing a periodic job!4
```

the code want job run every 2 minutes ,but it run every 2:30 minute , include the time the job costed.

Is this as expected? or Bug ？

Reproduce the example using schedule.every(2).minutes.do(job), schedule.run_pending(), and the 30-second sleep, then inspect how the scheduled job is run and rescheduled. Confirm whether the documented behavior should include job execution time; done means the timing semantics are clarified or the observed interval is corrected and verified against the example.

Written by the indexing model from the issue text.