Practise the standard verbs for managing Flink checkpoints and recovery.
0 / 5 completed
1 / 5
Fill in: 'We ___ checkpoints at a fixed interval so a Flink job can resume from recent state instead of reprocessing an entire stream's history after a failure.'
We 'trigger a checkpoint' — the standard, established Flink collocation for initiating a consistent state snapshot. The other options aren't the recognised term here.
2 / 5
Fill in: 'Setting the checkpoint interval far too long for a high-throughput job can ___ a large amount of already-processed data replayed after even a brief restart.'
We say an overly long interval will 'leave' a large replay window on restart — the standard, natural collocation for the resulting cost. The other options aren't idiomatic here.
3 / 5
Fill in: 'We ___ checkpoint duration closely, since a checkpoint that starts taking longer than the interval between them can eventually stall the whole pipeline.'
We 'monitor' duration — the standard collocation for ongoing observation of a checkpointing-health metric. The other options aren't idiomatic here.
4 / 5
Fill in: 'We ___ a job from a specific savepoint deliberately when deploying a schema change, so state migrates cleanly instead of being reconstructed from scratch.'
We 'restore a job' — the standard, established collocation for resuming execution from a saved state snapshot. The other options aren't the recognised term here.
5 / 5
Fill in: 'We ___ a recovery from a recent checkpoint in staging before trusting the same job's fault tolerance in production, rather than assuming it will just work.'
We 'test' a recovery — the standard, simple collocation for validating fault-tolerance behaviour ahead of production. The other options are less idiomatic here.