Version 0.1, free and open source

Run duplicate requests once, even when workers crash

TaskFused gives each piece of work a name. When the same request comes in again, from the same program or another one, it joins the operation already under way and gets the same result. Operations live in one local SQLite file, so a job whose worker died is picked up again under the same ID.

1 run

For 100 identical requests: one created the operation, 99 joined it, and it ran once

1.1 s

From killing a worker mid-job to the same operation finishing on a new one, with a one-second lease

26

Storage failures forced across 13 kinds of write; each one rolled back with nothing half-saved

Apache 2.0

License for all of the code

Measured on a two-core Linux cloud machine with SQLite flushing every change to disk. The test logs are on GitHub.

Guides and notes

All posts →

Questions or ideas?

If TaskFused nearly fits your project, or you'd like it to do something it doesn't yet, write in.

Contact