You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A document to track the directions from 0.3, replacing #26. Our mid- and long-term goals, their [priority], (asignee) and any sub-tasks.
Any help is welcome with mentoring available for most tasks!
Remaining enhancements from v0.3
Will be updated after prioritization discussion.
Client-side protocols
Replace capnp RPC and the current monitoring dashboard HTTP API with common protocol.
Part of #11 (more discussion there) but specific to the public API.
A document to track the directions from 0.3, replacing #26. Our mid- and long-term goals, their [priority], (asignee) and any sub-tasks.
Any help is welcome with mentoring available for most tasks!
Remaining enhancements from v0.3
Will be updated after prioritization discussion.
Client-side protocols
Replace capnp RPC and the current monitoring dashboard HTTP API with common protocol.
Part of #11 (more discussion there) but specific to the public API.
Improve the dashboard with more information and post-mortem analysis
Fix current bugs
Custom tasks (subworkers) in more languages
Easier deployment in the cloud
Packaging for easier deployment
Multiple options, priorities may vary. (@spirali)
Improve Python API
Pythonize the client API.
Improve testing infrastructure
rain start
and running on OpenStack, Exoscale, AWS. Does not have to be a part of CI (even for running locally). Depends on / part of Startup and control scripts for CloudStack and others #37.More real-world code examples
Lower priority, best based on real use-cases. Ideas: numpy subtasks, C++/Rust subworkers
Enhancements to revisit in the (not so distant) future
utils/bench/simple_task_scaling.py
. The results as of 0.2 are here.The text was updated successfully, but these errors were encountered: