Key Default Type Description
execution.batch.adaptive.auto-parallelism.avg-data-volume-per-task
16 mb MemorySize The average size of data volume to expect each task instance to process if jobmanager.scheduler has been set to AdaptiveBatch. Note that when data skew occurs or the decided parallelism reaches the execution.batch.adaptive.auto-parallelism.max-parallelism (due to too much data), the data actually processed by some tasks may far exceed this value.
execution.batch.adaptive.auto-parallelism.default-source-parallelism
(none) Integer The default parallelism of source vertices or the upper bound of source parallelism to set adaptively if jobmanager.scheduler has been set to AdaptiveBatch. Note that execution.batch.adaptive.auto-parallelism.max-parallelism will be used if this configuration is not configured. If execution.batch.adaptive.auto-parallelism.max-parallelism is not set either, then the default parallelism set via parallelism.default will be used instead.
execution.batch.adaptive.auto-parallelism.enabled
true Boolean If true, Flink will automatically decide the parallelism of operators in batch jobs.
execution.batch.adaptive.auto-parallelism.max-parallelism
128 Integer The upper bound of allowed parallelism to set adaptively if jobmanager.scheduler has been set to AdaptiveBatch
execution.batch.adaptive.auto-parallelism.min-parallelism
1 Integer The lower bound of allowed parallelism to set adaptively if jobmanager.scheduler has been set to AdaptiveBatch
execution.batch.job-recovery.enabled
false Boolean A flag to enable or disable the job recovery. If enabled, batch jobs can resume with previously generated intermediate results after job master restarts due to failures, thereby preserving the progress.
execution.batch.job-recovery.previous-worker.recovery.timeout
30 s Duration The timeout for a new job master to wait for the previous worker to reconnect.A reconnected worker will transmit the details of its produced intermediate results to the new job master, enabling the job master to reuse these results.
execution.batch.job-recovery.snapshot.min-pause
3 min Duration The minimal pause between snapshots taken by operator coordinator or other components. It is used to avoid performance degradation due to excessive snapshot frequency.
execution.batch.speculative.block-slow-node-duration
1 min Duration Controls how long an detected slow node should be blocked for.
execution.batch.speculative.enabled
false Boolean Controls whether to enable speculative execution.
execution.batch.speculative.max-concurrent-executions
2 Integer Controls the maximum number of execution attempts of each operator that can execute concurrently, including the original one and speculative ones.