Prefect Reviews & Product Details

It is quite easy to annotate existing Python code and convert it into a Prefect flow without changing a lot of code. The hybrid architecture which uses agents or workers allows you to have a lot of flexibility on where and how the code is executed. This also allows you to have isolated environments for every flow in a cleaner way than something like Airflow. Review collected by and hosted on G2.com.

Prefect 2 is still a young framework and it is still under construction so expect it to have a lot of constant updates and possible breaking changes. Its community is not as big as something like "Airflow" so it lacks a lot of connectors and also some of them are not frequently updated. Review collected by and hosted on G2.com.

I recently had the opportunity to adopt a Prefect centric ETL orchestration pattern and I have to say, it has been a game-changer. It was perfect for our situation where we had multiple locations with different setups, but wanted a common logging and deployment pattern.

What impressed me the most was how user-friendly Prefect is. It's much more straightforward compared to complex solutions like Argo Events, and or relatively mature products like Airflow. If youre in the python space working on ETL, you should be using this product.

But what really got me excited was Prefect's recent introduction of their AI tool called Marvin. It's fantastic for small "AI functions" and it allows us to leverage chatGPT results as a structured object type. This has added a whole new level of efficiency to our ETL processes cleaning up free text fields into structured categories.

Overall, I can confidently say that Prefect has exceeded our expectations. It's been a breeze to use and has greatly improved our ETL workflow. I highly recommend it to anyone in need of a reliable ETL orchestrator. Review collected by and hosted on G2.com.

There have been several breaking changes pushed over minor releases. Additionally they seem to have siliently stopped fixing bugs related to the "agents" in favor of the "workers" approach. The fact that there are known bugs with agents that will not be fixed in favor of the user adopting an entirely new pattern does get frustrating as a paying customer. Review collected by and hosted on G2.com.

Using Prefect has transformed my workflow management. Its hybrid execution model offers unparalleled flexibility, allowing local or distributed system runs. I particularly appreciate the dynamic pipeline construction, which adjusts based on intermediate results. The rich state system provides an insightful view of task execution, and the native Python API integrates seamlessly with my existing stack. Moreover, the UI dashboard, Prefect Cloud, is a game-changer for monitoring and visualizing flow runs. With its robust features and active open-source community, Prefect is an essential tool in my data toolkit. Review collected by and hosted on G2.com.

Setting up flows on the cloud felt intricate, and mastering the visualization tools in Prefect Cloud took time. Occasionally, its vast features were overwhelming, and the dynamic pipelines, though innovative, sometimes added complexity. Review collected by and hosted on G2.com.

- Nice scheduling platform that allow our Business department to grow their autonomy by scheduling their script to increase their productivity

- it allow us to keep the monitoring of their scheduling with our Flow center of Excellence.

- Our administrators can easily spot flows in errors and fix them before the Business figured it out

- I like the Support onboarding to help us customise our setup with AWS

- Prefect team allow us to adpat our setup with the new version

- It helped us to customise the helm chart to deploy our infrastructure with our kubernetes cluster on prem so we can orchestrate both in the cloud and on premise. Review collected by and hosted on G2.com.

- Implementation can be a bit tough depending on your company security rules to open access with your Cloud provider (AWS in our case) Review collected by and hosted on G2.com.

Prefect is a new-aged orchestration tool, designed for cloud databases. At its core, it is based on Python, which makes it a very flexible, stable, and cost-efficient tool to orchestrate your data pipelines. With Prefect open source, you can literally do magic to your data pipelines at no cost. It helps to build resilient data pipelines by running the flows on flexible infrastructure like local processes, Docker, Kubernetes, AWS, Azure containers, etc. Your Prefect code stays within your organization, with only flow run metadata pushed to the intuitive Prefect UI to gain better control and visibility on your data pipeline execution. With Prefect, you can build scalable, automated, and resilient pipelines using features like log retention, concurrency retries, and RBAC-based execution. Review collected by and hosted on G2.com.

Sometimes it feels short of examples/demo code while referring to Prefect documentation, but I hope that will improve over time. There is no downside to using Prefect if you have Python expertise on your side. Review collected by and hosted on G2.com.

Prefect powers Data Scientists to iterate on prod code development faster and gives engineers flexibility and scalability to host it in multiple places including on-prem and on cloud. Especially easy to use for data scientists using the decorators compared to cumbersome syntax which need practice like Airflow. Review collected by and hosted on G2.com.

The tool is ever-evolving so some of the desirable features are not available and need to wait for them to be implemented. Review collected by and hosted on G2.com.

Overall - I appreciate the simplicity Prefect provides and recommend it to all small-mid data teams. It allows myself to scale data pipelines effortlessly, ensuring full focus to be on the code itself. In the time I have been using Prefect it has matured significantly, documentation is in a much better state and currently extremely helpful. The open-source packages coupled with the 'blocks' system has saved me a ton of time (and provides a silky smooth local > prod workflow). Review collected by and hosted on G2.com.

Not many, if any, negatives to say about Prefect. My biggest pet-peeve is potentially the scattering of kinda unhelpful views in the UI (though they are flexible, but just give me summary and overviews first please). This has recently been improved with the addition of a "dashboard" view, providing at a glance statuses. Still it isn't perfect, I'm sure they will continue to improve though. Review collected by and hosted on G2.com.

I like that the product is a open source code and that they provide you with a service (Cloud) that solves all the boring stuff like IAM /RBAC for you. The flows and task python APIs are great. Simple to understand for data scientists. One of the best platforms for ETL. Review collected by and hosted on G2.com.

Some of the concepts are quite complex (Infrastructure block) and the documentation is not explaining them well. You can then find official Prefect YT videos that do better. Some of the solutions are a bit lacky/sloppy once you really use them. For example, Prefect AWS worker that spawns ECS tasks will do that in a way that it triggers AWS API rate limiting. Review collected by and hosted on G2.com.

Accelerated our product release with robust Orchestration capabilities and ability to generate top notch KPIs. Great for our managed service platform. Review collected by and hosted on G2.com.

The Downside is when customers want the same functionality and insist upon using other tools for Orchestration. Doable, but not as efficient nor robust. Review collected by and hosted on G2.com.

It has a great developer experience: I can develop my workflows in a jupyter notebook(including Colab). This allows me to do experiments before moving code that works in a Python script. Prefect blocks are another great feature which simplifies separating config from code. Review collected by and hosted on G2.com.

Some abstractions leak. I've had some challenges setting up a cloud-run push pool with access to on-prem DBs via a serverless VPC connector. The job template has an issue that I managed to fix after reading through Cloud Run REST API docs. Review collected by and hosted on G2.com.