Best Practices: Files Batch Deletion (delete_batch & delete_batch/check)

[dbjkr]

Hi there!

Just wondering what is the best practice to wait until batch operation (e.g. files batch deletion) is completed? Specifically, how often should I call /delete_batch/check endpoint (to know the progress or error/completion) in case I performing the operation from a backend? Since no time estimation provided for an operation (as far as I understand), should I check like just every 1s, 5s, 10s? Whether the check timeout better be incremental or constant?

And for the sake of completeness, how much such approach could be better comparing with sequential files deletion one by one? Do rate limits apply if make a lot of consequent requests that do not intersect? For example, my case is deleting about 100-200 files, each up to 5 MB.

Thanks!

[Greg-DB · Accepted solution]

We don't have any official documentation on how long you should wait between checking on jobs like this. A few seconds, say 5s, seems reasonable. Whether to use a constant or variable period is also up to you.

We don't provide firm guidance on this because how long any particular job takes depends on a few factors, which can vary over time, and per job and account. You may want to try a few, see what is typical for your use case, and customize accordingly.

There is a rate limiting system, as well as some amount of locking involved, so we do recommend using the batch calls whenever possible.