Preparation 2: Develop a de-escalation strategy for scheduling tasks and quality inspection
Develop a de-escalation strategy for task scheduling:
L0 and L1 provide exclusive monitoring to business email list ensure that tasks receive alarm notifications in a timely manner.
During the promotion period, when resources are tight, the platform will adopt a task delay extraction strategy and a task one-click delay strategy for L2 and L3.
When necessary, open green channels for L0 and L1 tasks to ensure the normal operation of tasks.
L0 and L1 task node resources are allocated preferentially.
For high-risk operations such as the modification of mission-critical attributes and mission disabling, the platform has different control strategies for different levels.
Develop a data quality degradation strategy:
When the execution time of the quality rule reaches 30 minutes, a reminder will be sent to the person in charge of the quality division and the associated dispatcher to confirm whether to intervene;
When the execution time of the quality rule reaches 60 minutes, the system automatically terminates the quality inspection, and the associated scheduling tasks are executed normally.
Preparation 3: Develop sealing management measures for scheduling tasks (create, copy, disable, rerun, etc.)
During the preparation period for the big promotion, if users create and copy tasks, because the security of new tasks cannot be guaranteed, there will be problems such as low performance, high resource occupation, and system stability. The following control measures: Prohibit the creation and copying of tasks, and require the approval of the person in charge of the secondary department.
The newly created task items need to be checked one by one, including:
The attention task period is hours and minutes;
Evaluate the impact of tasks on the system;
Check the where condition of the outbound task SQL;
It is recommended that users configure task monitoring and timeout.