Build monitoring mechanisms for Fivetran connectors

[plittlewood-rpt]

Background context The monitoring of existing Fivetran connectors is currently very light and somewhat inadequate (emails to individuals in the business). As adoption of the data warehouse increases, these connectors are more critical and should be monitored appropriated. We need to design a build an appropriate monitoring mechanism to allow for near real-time notification in the event of connector failures

Specification

• As this will follow on from #8486 consider there being multiple monitoring requirements including:
  • Monitoring of user actions (via the API built in the previous ticket). Display of audit logs (with optional searchability)
  • Connectors might be deliberately paused or stopped, so having a time based aspect to this is important (ie more urgent priority on connectors which have been paused through a user action and not restarted after x minutes, which could indicate an automation failure in a different service)
  • We will need an overall 'current' view of all connectors
• The Fivetran REST API can be used to retrieve connectors and understand the status of each.

[plittlewood-rpt]

I have reached out to Devops as there is a simpler route for this by using the external log connector to ship the logs out to Cloudwatch and then streamed into our existing Opensearch cluster for analysis and automated fault detection.

[plittlewood-rpt]

Dropping from sprint - waiting on DevOps. Internal ref ACP-2560