Need advice about which tool to choose?Ask the StackShare community!
Airflow vs Apache NiFi: What are the differences?
Developers describe Airflow as "A platform to programmaticaly author, schedule and monitor data pipelines, by Airbnb". Use Airflow to author workflows as directed acyclic graphs (DAGs) of tasks. The Airflow scheduler executes your tasks on an array of workers while following the specified dependencies. Rich command lines utilities makes performing complex surgeries on DAGs a snap. The rich user interface makes it easy to visualize pipelines running in production, monitor progress and troubleshoot issues when needed. On the other hand, Apache NiFi is detailed as "A reliable system to process and distribute data". An easy to use, powerful, and reliable system to process and distribute data. It supports powerful and scalable directed graphs of data routing, transformation, and system mediation logic.
Airflow can be classified as a tool in the "Workflow Manager" category, while Apache NiFi is grouped under "Stream Processing".
Some of the features offered by Airflow are:
- Dynamic: Airflow pipelines are configuration as code (Python), allowing for dynamic pipeline generation. This allows for writting code that instantiate pipelines dynamically.
- Extensible: Easily define your own operators, executors and extend the library so that it fits the level of abstraction that suits your environment.
- Elegant: Airflow pipelines are lean and explicit. Parameterizing your scripts is built in the core of Airflow using powerful Jinja templating engine.
On the other hand, Apache NiFi provides the following key features:
- Web-based user interface
- Highly configurable
- Data Provenance
Airflow is an open source tool with 13.3K GitHub stars and 4.91K GitHub forks. Here's a link to Airflow's open source repository on GitHub.
According to the StackShare community, Airflow has a broader approval, being mentioned in 98 company stacks & 162 developers stacks; compared to Apache NiFi, which is listed in 10 company stacks and 12 developer stacks.
I am looking for the best tool to orchestrate #ETL workflows in non-Hadoop environments, mainly for regression testing use cases. Would Airflow or Apache NiFi be a good fit for this purpose?
For example, I want to run an Informatica ETL job and then run an SQL task as a dependency, followed by another task from Jira. What tool is best suited to set up such a pipeline?
I have been using Airflow for more than 2 years now and haven't thought about moving to any other platform. Coming back to your requirements, Airflow fits pretty well. 1. It has an excellent way to manage dependent tasks using DAG (Direct Acyclic Graph), You can create a DAG with tasks and manage which task is dependent on which and Airflow takes care of running it or not running a task in case the parent task fails. 2. Integrations - The airflow community has implemented various integration to different cloud services, to Hadoop, spark a and as well as Jira. Though it doesn't have in-built integration for Informatica you can also run your own service in Airflow as a task (which can handle all Informatica related operations).
- It's very easy to find/monitor and manage Jobs/Pipelines as Airflow provides a great consolidated UI.
Hey Sathya! With Airflow, you are able to create custom hooks and operators to trigger various types of jobs. There may be ones that exist already for informatica, but I am unsure. Would be happy to connect to discuss further if you are interested. email@example.com
I am so confused. I need a tool that will allow me to go to about 10 different URLs to get a list of objects. Those object lists will be hundreds or thousands in length. I then need to get detailed data lists about each object. Those detailed data lists can have hundreds of elements that could be map/reduced somehow. My batch process dies sometimes halfway through which means hours of processing gone, i.e. time wasted. I need something like a directed graph that will keep results of successful data collection and allow me either pragmatically or manually to retry the failed ones some way (0 - forever) times. I want it to then process all the ones that have succeeded or been effectively ignored and load the data store with the aggregation of some couple thousand data-points. I know hitting this many endpoints is not a good practice but I can't put collectors on all the endpoints or anything like that. It is pretty much the only way to get the data.
For a non-streaming approach:
You could consider using more checkpoints throughout your spark jobs. Furthermore, you could consider separating your workload into multiple jobs with an intermittent data store (suggesting cassandra or you may choose based on your choice and availability) to store results , perform aggregations and store results of those.
Spark Job 1 - Fetch Data From 10 URLs and store data and metadata in a data store (cassandra) Spark Job 2..n - Check data store for unprocessed items and continue the aggregation
Alternatively for a streaming approach: Treating your data as stream might be useful also. Spark Streaming allows you to utilize a checkpoint interval - https://spark.apache.org/docs/latest/streaming-programming-guide.html#checkpointing
Pros of Airflow
- Task Dependency Management14
- Beautiful UI12
- Cluster of workers12
- Open source6
- Complex workflows5
- Good api3
- Apache project3
- Custom operators3
Pros of Apache NiFi
- Visual Data Flows using Directed Acyclic Graphs (DAGs)16
- Free (Open Source)8
- Reactive with back-pressure5
- Scalable horizontally as well as vertically5
- Fast prototyping4
- Bi-directional channels3
- Data provenance2
- Built-in graphical user interface2
- End-to-end security between all nodes2
- Can handle messages up to gigabytes in size2
- Hbase support1
- Kudu support1
- Hive support1
- Slack integration1
- Support for custom Processor in Java1
- Lot of articles1
- Lots of documentation1
Sign up to add or upvote prosMake informed product decisions
Cons of Airflow
- Running it on kubernetes cluster relatively complex2
- Open source - provides minimum or no support2
- Logical separation of DAGs is not straight forward1
- Observability is not great when the DAGs exceed 2501
Cons of Apache NiFi
- HA support is not full fledge2