TiSpark master component is deployed. Only one node of TiSpark master can be deployed. • tispark_workers: The configuration of the TiSpark worker instance. This configuration specifies the machines to which 12 tispark_workers tispark_workers specifies the machines to which the worker nodes of TiSpark are de- ployed. It also specifies the service configuration on each machine. tispark_workers is an array array element contains the following fields: • host: Specifies the machine to which the TiSpark workers are deployed. The field value is an IP address and is mandatory. • listen_host: When the machine

test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178 we use sysbench to simulate this workload. Specifically, run the following command to enable 10 workers to continuously write data to three tables, sbtest1, sbtest2, and sbtest3, with a total TPS not exceeding we use sysbench to simulate this workload. Specifically, run the following command to enable 10 workers to continuously write data to three tables, sbtest1, sbtest2, and sbtest3, with a total TPS not exceeding

test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178 we use sysbench to simulate this workload. Specifically, run the following command to enable 10 workers to continuously write data to three tables, sbtest1, sbtest2, and sbtest3, with a total TPS not exceeding schema version changes correspondingly and each version change is synchronized to other TiDB DDL workers (including BR). Therefore, when restoring a large number of tables, the serial execution implementation

Controls the number of encoding and decoding workers in the redo module. TiCDC flush- �→ worker �→ -num Newly added Controls the number of flushing workers in the redo module. 54 Configuration file point, you can consider increasing tidb_ttl_ delete_worker_count to increase the number of delete workers. For example: Figure 55: scan fast example 389 In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker_count �→ can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage

point, you can consider increasing tidb_ttl_ delete_worker_count to increase the number of delete workers. For example: Figure 55: scan fast example 402 In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker_count �→ can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178

point, you can consider increasing tidb_ttl_ delete_worker_count to increase the number of delete workers. For example: Figure 55: scan fast example 409 In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker_count �→ can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178

cgroup CPU and memory limits #16392 @pingandb 77 • Add CPU monitoring for Region workers and snapshot generation workers #16562 @Connor1996 • Add slow logs for peer and store messages #16600 @Connor1996 point, you can consider increasing tidb_ttl_delete_worker_count to increase the number of delete workers. For example: 417 Figure 55: scan fast example In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker �→ _count can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage

point, you can consider increasing tidb_ttl_delete_worker_count to increase the number of delete workers. For example: 388 Figure 55: scan fast example In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker �→ _count can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178

point, you can consider increasing tidb_ttl_delete_worker_count to increase the number of delete workers. For example: 394 Figure 55: scan fast example In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker �→ _count can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178

point, you can consider increasing tidb_ttl_delete_worker_count to increase the number of delete workers. For example: 397 Figure 55: scan fast example In contrast, if the scan worker is rarely in the tidb_ttl_scan_worker �→ _count can make the TTL task workload more balanced. Since too many TTL workers will cause a lot of pressure, you need to evaluate the CPU level of TiDB and the disk and CPU usage test is finished, the test summary results are printed. For example: Finished: 50 OLTP workers, 1 OLAP workers [Summary] DELIVERY - Takes(s): 3599.7, Count: 501795, TPM: 8363.9, Sum(ms): �→ 63905178