You’ve likely heard us talk about deploying code fast and recovering from incidents even faster. Well, we’ve got a host of product updates for you today that double down on these promises.
What’s new? Collaborate on incident recovery, new surface areas for manual interaction with workflows, and deploy your code on Prefect-managed compute for free.
Here’s the high level, read on for more:
We’re very excited about the new use cases that these features will unlock for you, especially with interactive workflows. Stay tuned in the days to come for deep dives on what’s now possible for you, and your workflows, with Prefect.
We’re introducing the ability to interact with a workflow during its runtime. What can you do with interactive workflows? A lot of surprising new things. Set up approvals for specific flows, create flows with manual, at-runtime inputs, or build flow to flow interactions.
To get set up: describe the data your flow needs, and Prefect will render a form with client-side validation and provide an API endpoint clients can use to submit input. Stay tuned for more on this feature, we’ll be sharing a deep dive next week.
You can now execute your code without having to configure or manage your own infrastructure with two new compute options.
First up: Prefect-managed compute is live in beta. Deploy your code and Prefect can execute it on your behalf at no charge. Free accounts get 10 hours of runtime monthly, and Pro accounts get 250 hours of runtime per month. Note that these are hard caps.
Use a Prefect-managed work pool to get into production with extreme speed or clean-up after your other flows fail, without any lock-in trap. Read more in the docs to get started.
Another infrastructure addition: Modal is now natively integrated with Prefect. Run your Prefect flows on Modal and enjoy crazy fast startup times and autoscaling for even the chunkiest of workloads, without any configuration needed.
Read more here.
When a workflow disruption occurs, you can now declare and manage incidents with Prefect. Assess impact, document actions, and resolve critical failure while keeping stakeholders informed.
Using Prefect incidents means you will detect outages earlier, recover faster, and document remediation steps for compliance and stakeholder management. Incidents can be declared manually or automatically (for instance, based on flow run performance metrics via an automation).
Metrics are aggregated views of events running through Prefect. Monitor and automatically take action based on metrics such as average workflow lateness, duration, and success.
Want to establish an SLA for a particular set of flows? Configure an automation that alerts and takes automated action (such as declaring an incident or sending a notification) when your average lateness hits your SLA. Read more about using metrics in automation triggers here.
Follow along for deep dives and real use cases over the next couple of weeks demonstrating each of these new releases. We can’t wait to see how you customize Prefect with these new tools.
Want to follow along? Sign up for our newsletter or try out Prefect for yourself!
