List Flink savepoint operations

Question

There are 2 savepoint related REST APIs:

POST /jobs/:jobid/savepoints which triggers an async savepoint operation and returns a trigger ID.
GET /jobs/:jobid/savepoints/:triggerid which returns the status of the operation by trigger ID.

But in a distributed system, what if the operation is successfully initiated, but the response is lost (e.g., because of network issue), then the trigger ID is lost, there is not way to figure out the status of the operation.

Wondering is there a way to list Flink savepoint operations? so that I can get the trigger ID and then get the operation status.

You could get a checkpoint history (which also contains the savepoints). But the history contains only the last 10 checkpoints (no clue if you can change this). — TobiSH
There is an attribute is_savepoint at each checkpoint in the json response: ci.apache.org/projects/flink/flink-docs-stable/monitoring/… — TobiSH

TobiSH TobiSH · Accepted Answer · 2019-10-22T08:14:37

The only way I am aware of (beside iterating over the savepoints dir) is using the rest interface to get a checkpoint history and filter out the savepoints: https://ci.apache.org/projects/flink/flink-docs-stable/monitoring/rest_api.html#jobs-jobid-checkpoints

Downside: You only get the last 10 checkpoints :-(

List Flink savepoint operations

1 Answers