[Proposal] Forecasting Dev Ops Metrics and Financial/Inventory Metrics

The OpenSearch project wants to offer Forecasting as a tool for the community. There are a few different ways that we think that Forecasting would be helpful.

  1. Augmentation for visuals on Dashboards to help developer operations in live troubleshooting
  2. Integrating with alerting to gently let users/systems know when a service may be trending in the wrong direction.
  3. Performing forecasting in the traditional business/finance sense on large amounts of data which includes seasonality (e.g. predicting what revenue may look like or inventory levels based on prior data and trends)
  4. An integration to make it easier to manage Forecasting functionality via the API

Are there use cases and nomenclature in our community that we are missing? We would love your feedback.

1 Like

Thanks @jdbright for posting here! Hope the community has a lot of input for you

Quick update. The project can really use some feedback on what the community values.

The team is focusing on the Dev Ops persona first. A user will be able to use Dashboards to augment their time series charts with a trending view into the future. Similarly, the team wants to add forecasting to alerts to let people know when a given metric (e.g., CPU, memory, 500 errors) is trending in the wrong direction ahead of a formal alert being fired.

Community members, do you see both pieces of functionality as equal or is one higher priority than the other?