price below 100
We have introduced a new pricing https://www.windmill.dev/pricing
Instead of vCPUs, we are now basing pricing on compute units (a 2GB worker used for a full month). Workers have a size (small, standard, large) based on their memory limit (<=1, (1-2], >2) and each worker size gets a compute unit (0.5, 1, 2)
With this new approach, you can now get rid of setting vCPU limits which is an anti-pattern in favor of only setting them as requests not to limit unnecessarily the compute power of your workers...
**Many UX improvements for the flow
Many UX improvements for the flow builder
We have lots of nice changes we haven't announced but here is one bulk changelog for the flow editor:
- If you write a static string that looks like an expr we will offer to make it an expr as a TAB
- you can now connect to nodes directly using a plug system similar to the nodes for apps...
As we prepare for auto-scaling (EE), we
As we prepare for auto-scaling (EE), we needed more metrics to be held internally such as occupancy_rates at different timelines:
Occupancy rates at different timelines
You can now see occupancy rates for every worker in the 15s/5min/30mins/ever timelines to give you a glimpse of how busy your workers are right now...
**Major**: See service logs directly in
Major: See service logs directly in windmill:
See the logs of any workers or any servers in the service logs of the search modal. Log count is seen in the mini-graph and error count is in red...
recover error
New option for flow error handlers: recover the flow if
recover: true is returned as part of the error handler result, also we standardized step_id to be in the default template
Retry #N
Retries are now easier to understand in the flow status viewer:
You can see their numbers as well as the detail of the failed jobs at a glance...
