Hi! Please make sure to always cc the user mailing list. Otherwise other people will not see your mail.
Can the checkpoint be manual trigger too? No you can't. Checkpoint can only be triggered by time. This is one of the main difference between checkpoint and savepoint [1]. What's your use case by the way? Why would you like to trigger checkpoint according to the number of messages consumed? [1] https://nightlies.apache.org/flink/flink-docs-release-1.14/docs/ops/state/savepoints/ ZhangJian He <shoot...@gmail.com> 于2021年11月15日周一 下午11:10写道： > Really Thank you. I know I can trigger savepoint through `/bin/flink > savepoint $JOB_ID /tmp/flink-savepoints` > Can the checkpoint be manual trigger too? > > Caizhi Weng <tsreape...@gmail.com> 于2021年11月15日周一 上午9:49写道： > >> Hi! >> >> Flink currently does not have an automatic way to achieve this. However >> you can use an external system to monitor on Kafka metrics [1] (for example >> currently consumed offsets) and trigger savepoints when certain conditions >> are met. >> >> [1] >> https://nightlies.apache.org/flink/flink-docs-master/docs/connectors/datastream/kafka//#monitoring >> >> ZhangJian He <shoot...@gmail.com> 于2021年11月14日周日 下午3:11写道： >> >>> Hello, I am new to flink. >>> I want to know if I can generate savepoint according to the consuming >>> message numbers(in Pulsar/Kafka) >>> >>> >>> Thanks for your help >>> ZhangJian He >>> >>>
