The entire ETL jobs written in python are currently scheduled in Glue with the s3 trigger attached , I should say it works so seamlessly !

The AWS glue give you the power to integrate/migrate any data sources and build the semantic layer for reporting and other data analytical purposes . The crawler of Glue works brilliantly well to scan the source files to perform initial validations to move the data across the next stage .

The scheduler of the glue could have been better , Since its currently misses out in lot of features . Even the job dependencies cannot be set at present .

A good thing about it is that you don't have to worry about managing the servers, scaling or anything related to the infrastructure.

The good thing can sometimes be bad too, AWS Glue provides pre-built templates, there are limits to the customization that can be done.

We ultimately found AWS Glue quite complex to achieve the things we wanted, and ended up switching to a different product.

AWS Glue offers a lot of functionality and fits in well with other AWS products. In our case, we wanted to migrate data from a bunch of different AWS RDS databases and products to a data warehouse. AWS DMS got us only so far, whereas Glue looked like it would help us take deltas in S3 buckets and merge them together and transform them. So, in that regard, it's great as an all-in-AWS solution that can work with AWS S3, DMS, QuickSight, IAM, etc.

It's not easy to set up or debug. In fact, you'll end up scratching your head many times as you try and work out why things aren't working. You need to really look at it hard and implement logging if you're going to work it out. However, frustratingly, only a small amount of its power is available within the GUI. For instance, you can do parquet file formats in the GUI and need to use JSON or the CLI.

It is an ETL tool in AWS that you can use for moving the data between different databases and services like S3, data archiving, and transformation. It has got an inbuilt spark code editor where you can write your own customized code.

It gets complicated very quickly and the whole process of using this tool for complex tasks requires immense effort and time. In terms of pricing, it can get expensive based on the performance you are expecting. Debugging the issues that are related to permissions while using this service is a nightmare and you end up giving full permissions.

Yo y mi organización llevamos algunos años utilizando este software para el análisis, integración y limpieza de datos de nuestra empresa y puedo decir de primera mano que es perfecta y permite ahorrar meses de trabajo y resultados increíbles.

Este software sin duda es mas potente en cuanto análisis, integración de datos y limpieza, permitiendo ahorrar mucho tiempo y observar las tareas perfectamente dando lugar a resultados fantásticos para mi trabajo.

No tiene nada negativo, todos los aspectos del software son imprescindibles y beneficiosos.

Since AWS Glue comes easily integrable with other AWS Tools, it's easy to set up and looks like a clear choice when dealing with consumption of large amount of data. Also, monitoring and dashboards are easy to maintain for AWS Glue.

Rates of data consumption for AWS Glue cannot be easily managed, which leads to data loss and corruption. Also, it can only process data in batches but not real time data.

AWS glue is a wonderful ETL apparatus by AWS. It is a server less and end client don't have to deal with any infrastructure simply foster the pipeline utilizing paste studio or custom python prearrange and execute it exceptionally quick simple to learn and easy to execute .

If the at first beginning up an ideal opportunity for cluster turn up could be diminished it would be an extraordinary help to deal with.

Complexité d'utilisation

AWS Glue est un ETL(Extract, Transform, Load), gérer par amazone qui est un connecteur de plusieurs source de données

Dépendances à Amazone, coût de déploiement élevé

It's a decent ETL tool for the price. It can handle basic transformations and it's easy to use. Anyone with a desire to learn it could do so easily.

It's very easy to use. It seems like it's meant for first time ETL engineers. There is enough documentation to learn it on your own, and the costs isn't so bad.

It's too simplistic. It doesn't handle complex transformations very well.