AWS AppFlow trickles out, aiming to tighten the net around SaaS apps


In a bid to lure more devs onto its company’s services, AWS has pushed out AppFlow, claiming it will lower the hurdle of connecting software-as-a-service apps to analytics measures.

The new product is meant to free developers from having to manage connectors or write code in order to transfer data to a number of services for processing or storage tasks. AppFlow can run up to 100GB of data per set up connection and is marketed to those realising automatic data backup tasks, event analysis, reporting, and synchronisation between instances.

To set up a so-called “flow”, AWS customers can choose a data source from a selection of supported SaaS applications such as Slack, ServiceNow, Zendesk, or Salesforce. Once authenticated, those can then be connected to an Amazon S3 bucket, Amazon Redshift, Salesforce, or data warehousing product Snowflake.

Source data fields can be mapped either manually or via a .csv file to their source equivalent. AppFlow also offers transformations like masking (for example to keep some information private) and merging, which can be applied during transfer; filters to, for example, limit the data transferred to the newest additions only; and validation options to make sure fields with little information aren’t transmitted.

Flows are bidirectional, meaning that sources can become destinations and the other way round, if set up that way. To trigger them, users can either set a specific time, define the occurrence of certain events as starting points for an operation, or click the “run” button in the AppFlow user interface. 

Data transfer is always encrypted, the company assures, though users who feel like their data is better off on AWS’ network are free to opt for sources that have AWS PrivateLink enabled, which “creates and configures private endpoints so your data remains private by default”. And since it’s an AWS service, admins are free to work with existing identity and access management policies to make sure not every change has to be sanctioned by the infra team.

Pricing depends on the number of flows run and the amount of data processed, which again depends a bit on the region the services are used in and tends to be slightly higher in the Asia Pacific areas. Users will also have to take into account the fees for data requests and storage when combining AppFlow with S3 and the cost of AWS Key Management for encrypting access tokens and transit data. 

Currently, flows are restricted to 10 million per account and month, with one flow costing $0.001. Data processing costs are $0.08 per GB in most regions, except for cases where destinations aren’t hosted by AWS and AWS PrivateLink isn’t integrated, where users will have to pay $0.136 per GB when more than 1GB is processed.