The ability to schedule tasks at scale and then have clear observability into how they’re progressing is really valuable. Our tasks run smoothly even at larger scale, without issues. The interfaces also give us a solid overview of how our tasks are running, which makes it easier to monitor everything. On top of that, it includes a lot of out-of-the-box features that help streamline the full workflow. Review collected by and hosted on G2.com.

The documentation is very poor for most use cases, and there were several times when we had to figure things out on our own. On top of that, support is limited to the self-hosted version, and the community wasn’t that helpful either. We ended up going for a paid plan, but for our use case it wasn’t ideal, and the cost didn’t feel worth it. Because of that, we ultimately managed to make it work with the self-hosted version.

I don’t think the product is very popular yet, but when it becomes more widely used, it will grow a good following. Review collected by and hosted on G2.com.

It is easy to stand up a flow, get UI benefits for orchestration, and test and debug when flows fail. Developing prefect flows in v2/v3 of prefect is also much more straightforward than it was on prefect v1. Review collected by and hosted on G2.com.

When getting into really large scale runs, it can sometimes choke. I feel the solutions their team gives us are kind of bandaid hacks to things their platform should handle well. Some of those orchestration elements like automatic clean up of stuck jobs is something I think an orchestration tool should provide out of the box. Review collected by and hosted on G2.com.

Prefect is one of the low costing and best orchestration tool in market. We can integrate lots of platforms with prefect. Databricks and snowflake jobs can be ran through prefect. Review collected by and hosted on G2.com.

Need Python knowledge to work on prefect and need to have some more features where Airflow is ahead Review collected by and hosted on G2.com.

Prefect is one of the best orchestration tools. Using the source git repo, we can customize integration our own way. Easy to use even for beginners using local setup. Knowledge of Python is sufficient. Review collected by and hosted on G2.com.

Other major competitors to Prefect have pre-built hooks and connectors where Prefect is behind because we should develop the connectors. Review collected by and hosted on G2.com.

The UI is the best part. All of our production processes, prefect or not, send error messages to slack. Our team members rotate being on call, so if an error message persists in slack, whoever is on call has to investigate. With most of our processes, the on call person has to get out their laptop, track down logs, etc. If they are in the middle of something, such as dinner, friends, etc, they have to go back to their laptop to handle things. With prefect, I am easily able to go on my phone when a prefect error comes in, log into prefect, and take care of everything I need to right there. An ios app for prefect would make this even better than it already is. Review collected by and hosted on G2.com.

Documentation is terrible. Not enough examples, especially with the more complicated things like work pools and blocks. Any examples that are there are way to simple to actually help you implement it on something real. The great design of the UI give you better clues as to how things work than the documentation does. Review collected by and hosted on G2.com.

Easy to understand framework and concepts for new DataOps position, comparing to Airflow

Supportive and helpful through Slack channel, typically get response within a hour Review collected by and hosted on G2.com.

Nothing so far, we are happy with Prefect Free Versions Review collected by and hosted on G2.com.

The thing our team has enjoyed the best about the prefect is how easy it is to convert any python code into a working and automated pipeline via the prefect decorators. We were able to migrate our cloud function workflows into prefect in just a couple of days. The declarative deployments yaml file is also easy to understand and when used in our CI/CD pipelines. Review collected by and hosted on G2.com.

I wish that prefect had more integrations but I do understand that it is growing and the team is working hard to grow this area. One thing that I wish prefect cloud had a native integration with Kafka and integrations with DataHub. Review collected by and hosted on G2.com.

Prefect is all about taking the idea of having scheduled Python scripts and easily and expertly tacking on all the things that make an engineer's life easier: UI-based scheduling, parameterization, flow visualization, detailed logging, exception handling, notifications, webhooks, containerization and easy deployments, concurrency. It's like a pure (and massive) value-add system with almost no need to make concessions in how you like to connect to data, transform data, or deliver data. As such I think it's especially powerful when you have a lot of non-standard ETL pipelines going in all directions (i.e. a mountain of tech debt) and you want to clean things up without being forced into restrictive design decisions.

Prefect is really flexible, not just in terms of how you deploy the execution agents, but also in terms of how you want to structure your code and what Prefect features you want to use heavily or lightly. There's no limitation on data sources or integrations: it just fits into whatever you're already doing with Python. It doesn't force you into any mold beyond just using Python, although the concepts of tasks, and the way Prefect handles concurrency, is super helpful if you want to make use of these things. It makes it easy to containerize Python scripts and make every aspect of a data flow deployment modular and flexible. And once you have your settings the way you like them, deploying changes or new flows is easy.

We run about 200 scheduled Prefect flows a day, sending and receiving data between over 50 different systems at our university. Prefect has been a lifesaver. Documentation, support, and the Prefect Slack space have all been super helpful. Review collected by and hosted on G2.com.

I can't complain hardly at all about Prefect. Having said that, I'm an experienced engineer who likes a lot of control. If you're already writing scheduled Python scripts to move data, Prefect will be a great fit, I think. But it isn't designed to address questions like how to extract data from X system or load it into Y system or how to transform data. It's about simplifying all the scaffolding around the guts of your Python code so you can focus on doing what you need to do. Review collected by and hosted on G2.com.

It is pythonic and therefore easy to implement. Version 2.0 has great features, like deploy. This allows us to develop remotely, collaboratively but in an orderly and simple way. Once you understand the essentials, i.e. flows, tasks, work-pool, workers and deployment, everything becomes very simple. It has a great community, very active. In general, the problems I've had already have several comments and solutions on Slack. Review collected by and hosted on G2.com.

We had to read and test for a couple of days to get the development environment up and running and connect to the cloud. In the documentation, more examples are missing, such as the connection with Gitlab (there are only examples with Github) and there are a couple of differences that we solved with the community. Review collected by and hosted on G2.com.

I've followed Prefect since version 1, and have happily observed as my suggestions on their (very responsive) Slack-channel have made it into production.

Very little initial code is required to register a flow, subflows and subtasks, but the options are vast. I do relatively simple ETL-jobs, but find myself eliminating a lot of the manual checks as I can trigger flows based on events in others.

The API is decently documented.

Elegant retries, flow-naming and options for task-runners.

If your are like me and found Airflow daunting - this is a really good choice. Review collected by and hosted on G2.com.

It's in rapid development, so pay attention to the release logs. As an example, I used to deploy using Python objects, but now there preferred option is the prefect.yaml file.

Some of the development is not always reflected in the documentation, but they are catching up.

The Flow-view can be a bit cluttered if you have some jobs that run for a few minutes and others for hours. Review collected by and hosted on G2.com.