Amazon EMR Reviews & Product Details

What I like best about Amazon EMR is its ability to process large scale data workloads using frameworks like Spark, Hadoop, and Hive without manually setting up complex big data infrastructure. It provides flexibility while integrating well with the Amazon AWS ecosystem. The flexibility of choosing frameworks while leveraging cloud scalability is very helpful. It also provides strong processing capability. Review collected by and hosted on G2.com.

Cluster configuration and optimization can become complex, especially for large production workloads. Cost management also requires attention because poorly configured clusters can lead to unnecessary compute usage. Also, there is no serverless model in Amazon EMR as Dataproc serverless in GCP, which I don't like. Review collected by and hosted on G2.com.

I currently use Amazon EMR to run Spark ETL workloads and orchestrate large-scale data processing pipelines. EMR helps me execute jobs, optimize Spark performance, and analyze execution time. The best part is that it seamlessly integrates with S3 and Airflow, which I like the most. Review collected by and hosted on G2.com.

I don’t have any major dislikes about Amazon EMR. Overall, it’s a good product and has worked well for me. Review collected by and hosted on G2.com.

EMR helps us work with big data in a faster and easier way. The UI is simple for starting jobs, checking progress, and seeing what’s happening, including reviewing logs. It also works well with other AWS services like CloudWatch. On pricing, it can save money because we only use what we need instead of running everything all the time. Support and onboarding are straightforward; if you’re already using AWS, it’s ready to use. It also has smart features like scaling and tuning, which make it easier to manage. Gives good performance. Review collected by and hosted on G2.com.

Amazon EMR is helpful, but it’s hard to learn at first. The screens and setup can feel confusing, and costs can go up if they aren’t managed correctly. Review collected by and hosted on G2.com.

I like that Amazon EMR is easy to use, robust, and reliable. It's great that I don't have to worry about the infrastructure and the servers that my distributed load runs on. It's easy to copy my PySpark files onto the EMR cluster and run them. The auto-scaling feature is very helpful and having auto termination is convenient as well. Review collected by and hosted on G2.com.

Sometimes there are edge cases where we have problems with the cluster. Sometimes we lose our spot nodes, so jobs fail. It can be a bit difficult to debug these edge cases, and that requires opening a ticket with Amazon support. Review collected by and hosted on G2.com.

1. Launch complex frameworks like Spark or Hive in minutes.

2. Keep data in cheap S3 storage and only pay for compute when needed.

3. Pricing up to save up to 90% on processing.

4. Built-in security feature is helpful.

5. UI is user-friendly and easy to understand. Review collected by and hosted on G2.com.

1. It can take several minutes.|

2. EMR sometimes runs slightly older versions of open-source tools compared to the very latest releases.

3. It still requires a good amount of technical knowledge to configure and optimize properly. Review collected by and hosted on G2.com.

EMR is easy to run single job for testing and it gives exact log error Review collected by and hosted on G2.com.

Sometimes it fails because of scaling up/down. So we have to tune it manually Review collected by and hosted on G2.com.

EMR we are using for running business logic on large business data received from various sources and third party vendors Review collected by and hosted on G2.com.

Auto scaling for task and core nodes is slow and takes more than 15 minutes normally which causes failure of running jobs due to lack of resources on cluster. Review collected by and hosted on G2.com.

It is very easy to launch or clone EMR cluster. And EMR provides very easy scaling capabilities based on containers, cpu, spot instances, usage of instance fleet or instance groups. And EMR supports many of the widely used applications like Spark, Hive, Hadoop, Trino, Presto, Ranger, Flink etc Review collected by and hosted on G2.com.

Working with Spot instances on EMR is slightly complicated during unavailability of spot instances when you need to use instances on one particular availability zone. Many solutions like databricks provide fallback which are even more easy to use Review collected by and hosted on G2.com.

Amazon EMR is a much more powerful product to deploy big data solutions on top of Spark, Flink, scoop, etc. It is very is to configure and provides a nice UI that helps a lot in debugging spark jobs. Apart from that, from an observability point of view, EMR dumps all logs to S3 and the cloud watch which eventually helps developers to debug memory issues in the cluster Review collected by and hosted on G2.com.

I don't like the notebook interface it provides, it doesn't have features like auto-completion etc. Review collected by and hosted on G2.com.

Great User Experience and User Interface

Faster

More scalable

Can Automate Easily

Read input from multiple sources

Write output to multiple sources

Accepted different types of programming language Review collected by and hosted on G2.com.

Booting up takes time that needs more patience

great tool but expensive that is one of the main disadvantage

Other than that nothing, everything looks great, I am using it every day. Review collected by and hosted on G2.com.