Release Notes – 0.22.1

Hi all! It's been a while since our last announcement, we've been pretty busy and we are now at version 0.22.1!

Our new Frontend and Infrastructure team members have started contributing off the bat, so we've been able to add backend, frontend and infrastructure improvements that provide new functionality, make the framework easier to use, and make users more productive.

There are loads of new improvements and fixes, too many to list, so here is a sample.

Use our Helm chart to deploy Sematic to your Kubernetes cluster

Simply prepare your values.yaml file and off you go. Read documentation here.

Re-run pipelines from any step

Expensive pipelines that fail after 99% due to a corner-case bug are a bummer. You can now patch your bug and re-run the pipeline with the fix, without re-computing all the previously-successful runs.

Expand shared memory on Kubernetes workers

If you use computationally-intensive libraries like PyTorch that can use the shared memory partition for better performance, now you can configure the worker pods to expand this partition for faster execution time.

Deep-linking of tabs and runs in the Dashboard

No longer do you have to manually seek between runs and tabs; now the Dashboard URLs point to individual runs and tabs so you can bookmark or share a page with your teammates.

Improved log scrolling

Scrolling down makes it visually explicit that more logs will be fetched, and the performance of log collection has been improved.

Kubernetes pod errors exposed in dashboard

Infrastructure failures are tricky to debug, so now we're collecting error state information and exposing it in the log panel, alongside user code errors.

Welcome new contributors!

We are thrilled to welcome @kamalesh0406, @KatkaG, @materight, @erikcek as new contributors. Thank you for your contributions!

See the complete changelog for more details.

Also, check out our revamped technical documentation and our company website!

‍