[FLINK-20612] Add benchmarks for scheduler

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.13.0
• Fix Version/s: 1.13.0
• Component/s: Benchmarks, Runtime / Coordination
• Labels: pull-request-available
• Assignee: Zhilong Hong
• Reporter: Zhilong Hong
• Created: 15/Dec/20 10:27
• Updated: 23/Mar/21 03:27
• Resolved: 23/Feb/21 08:36

Description

With Flink 1.12, we failed to run large-scale jobs on our cluster. When we were trying to run the jobs, we met the exceptions like out of heap memory, taskmanager heartbeat timeout, and etc. We increased the size of heap memory and extended the heartbeat timeout, the job still failed. After the troubleshooting, we found that there are some performance bottlenecks in the jobmaster. These bottlenecks are highly related to the complexity of the topology.

We implemented several benchmarks on these bottlenecks based on flink-benchmark. The topology of the benchmarks is a simple graph, which consists of only two vertices: one source vertex and one sink vertex. They are both connected with all-to-all blocking edges. The parallelisms of the vertices are both 8000. The execution mode is batch. The results of the benchmarks are illustrated below:

Table 1: The result of benchmarks on bottlenecks in the jobmaster

	Time spent
Build topology	45725.466 ms
Init scheduling strategy	38960.602 ms
Deploy tasks	17472.884 ms
Calculate failover region to restart	12960.912 ms

We'd like to propose these benchmarks for procedures related to the scheduler. There are three main benefits:

1. They help us to understand the current status of task deployment performance and locate where the bottleneck is.
2. We can use the benchmarks to evaluate the optimization in the future.
3. As we run the benchmarks daily, they will help us to trace how the performance changes and locate the commit that introduces the performance regression if there is any.

In the first version of the benchmarks, we mainly focus on the procedures we mentioned above. The methods corresponding to the procedures are:

1. Building topology: ExecutionGraph#attachJobGraph
2. Initializing scheduling strategies: PipelinedRegionSchedulingStrategy#init
3. Deploying tasks: Execution#deploy
4. Calculating failover regions: RestartPipelinedRegionFailoverStrategy#getTasksNeedingRestart

In the benchmarks, the topology consists of two vertices: source -> sink. They are connected with all-to-all edges. The result partition type (PIPELINED and BLOCKING) should be considered separately.

Issue Links

• is duplicated by: FLINK-10320 Introduce JobMaster schedule micro-benchmark
• relates to: FLINK-21110 Optimize scheduler performance for large-scale jobs
• relates to: FLINK-21731 Add benchmarks for DefaultScheduler's creation, scheduling and deploying
• links to: GitHub Pull Request #7