import pandas as pd
support_log = pd.read_csv('/datasets/support_log.csv')
support_log_grouped = support_log.groupby('type_id').count()

def alert_group(messages):
    if messages <= 300 :
        return 'средний'
    if messages <= 500:
        return 'высокий'
    return 'критичный'

support_log_grouped['alert_group'] = support_log_grouped['user_id'].apply(alert_group)
print(support_log_grouped.groupby('alert_group')['user_id'].sum())