Creating an RTMP push input - AWS Elemental MediaLive

Creating an RTMP push input

Create your input before you create the channel that ingests the input.

For information about the types of RTMP inputs that MediaLive supports, see Reference: Supported input containers and codecs.

You can set up an RTMP input as single-class or a standard-class input. You should have already read Implementing pipeline redundancy, to decide the class of channel and class of input that you want.

To create an RTMP push input

  1. You should have already arranged with the video content provider to set up the upstream system for your content. Make sure that the operator of the upstream system gives you the following information:

    • The sets of IP addresses where the source or sources for the content will appear on the public network. You need this information to create the input security group that you attach to the RTMP input. For example:

      • For one source:,,

      • For the other source:,,

    • The application name and application instance for the source content. (The application instance is also known as the stream or stream key.) You need this information to create the RTMP input. For example:

      Application name: live, and instance name curling

  2. Open the MediaLive console at

  3. In the navigation pane, choose Inputs.

  4. On the Inputs page, choose Create input.

  5. Complete the Input details section:

    • Input name – enter a name.

    • Input type – choose RTMP (push).

  6. In the Network mode section, choose Public.

  7. In the Input security group section, specify the group to attach to this push input. You can choose an existing group, or you can create a group. The security group must allow the public network IP addresses to push to MediaLive. Following from the example in step 1, it must allow these addresses:,,,,,

    For more information about security groups, see Working with input security groups.

  8. In the Channel and input class section, choose the class for this input:



    For more information, see Implementing pipeline redundancy.

  9. In the Input destinations section, in the Destination section, enter the application names and application instances you previously obtained:

    • If the input is a standard-class input, complete both fields, to specify two sources.

    • If the input is a single-class input, complete the first field with the information that you obtained and leave the second field empty.

    For example:

    Application name: live

    Application instance: curling

  10. In the Tags section, create tags if you want to associate tags with this input. For more information, see Tagging AWS Elemental MediaLive resources.

  11. Choose Create.

    MediaLive creates the input and automatically creates two endpoints on that input. The endpoints include the application name, the application instance, and the port 1935. For example:

    Note that the IP addresses are addresses that MediaLive creates. They aren't the public addresses that you used in the security group. For a diagram that shows the role of all the IP addresses, see Result of this procedure in the section about setting up an RTMP push source.

    MediaLive always creates two endpoints:

    • If you will set up the channel as a standard channel, both endpoints will be used.

    • If you will set up the channel as a single-pipeline channel, only the first endpoint will be used. MediaLive won't expect to receive content at the second endpoint.

  12. Provide the upstream system with the following information:

    • If you will set up the channel as a standard channel, provide both locations. The upstream system must push the video streams to these locations.

    • If you will set up the channel as a single-pipeline channel, provide only the first location. The upstream system must push its one stream to this location.

    For example, provide these addresses:

Result of this procedure

As a result of this setup, an RTMP push input exists that specifies two URLs. These URLs are fixed for the lifetime of the input, regardless of changes that occur (such as modifying other information in the input, or attaching the input to a different channel).

The upstream system pushes the source content to these endpoints.

Keep in mind that with a push input, the upstream system must be pushing the video source to the input when you start the channel. The upstream system does not need to be pushing before then.

For a description of this setup that includes a diagram, see Result of this procedure in the section about setting up an RTMP push source.