Add telemetry

[TiganeteaRobert]

Addition of telemetry to the Stream Replicator so that we can start to understand some basic information about open source usage to help guide us on where we can improve our OSS:

• Application: so that we know which apps are most commonly used by open source users
• App version: so we can start to associate value to a release, and understand how many open source user are upgrading / which versions of our apps are in most use
• Heartbeat event: tells us how long an app is running for, so we can understand usage over time
• User provided id: an optional input, but when populated allows us to tie events together across applications so that we can get an idea of which applications are being used together
• Source & Target: an optional set of inputs, when populated it allows us to gather information on the usage of sources and targets, making it easier for us to benchmark their individual popularity.
• Filter: an optional input, when populated it allows us to understand better what filters are used by the users, making it easier for us to understand the most popular scenarios that implement the Stream Replicator.

Acceptance Criteria:

• The schemas for the telemetry events & entity are published to Iglu central

• Telemetry will be opt-out, and documentation on docs.snowplowanalytics.com describes how to disable this feature

• The configuration examples must include an example of disabling the feature

• It must send the following event on a regular basis (hourly):

  • application_heartbeat

• It must send the following fields in the context:

  • Application name
  • Application version
  • Auto generated ID
  • Cloud
  • Region

• It must be possible for users to provide the following field, but this is optional:

  • User provided ID
  • Source
  • Target
  • Filter

• The following fields in the context are optional and will only be provided when app is deployed using the quick start module (will be passed as environment variable):

  • Module version
  • Module name

Note: we will always be transparent about what we are tracking and why. You can find more about our telemetry principles on this page.

[colmsnowplow]

This is still open

[alexanderdean]

This implementation is really excellent. Can you clarify what is the User provided id?

[colmsnowplow]

Same thing as in the Open Source Quick start - users can optionally provide an ID to identify themselves to us.