This is the first iteration described in https://gitlab.com/gitlab-org/gitlab/-/merge_requests/131400: Getting forecast **data** for Dedicated tenants. As described in the Blueprint, we're going to setup Tamland to run inside the Dedicated environment for each tenant. It'll run regularly and connect the tenant-local Prometheus instance to get metrics data. This data is going to get cached in a S3 bucket. The results of one execution is a complete set of forecasting data for all components specified in Tamland's manifest. The resulting dataset is going to be stored in a S3 bucket, too. The manifest for Tamland (for .com) comes from the `runbooks` repository and is being generated from the metrics-catalog. We aim to find a similar method for generating the manifest for a tenant (detailed in https://gitlab.com/gitlab-com/gl-infra/scalability/-/issues/2272). The tenant-local automation around Tamland needs to grab the Tamland manifest from a to be specified location (likely a GitLab project on the ops instance, comparable to [this one](https://ops.gitlab.net/gitlab-com/scalability/tamland-gprd)). Definition of Done: - [x] Tamland's latest version executes within all tenant environments daily - [x] The manifest gets generated automatically from the metrics catalog - [x] Resulting forecast data is being stored in the S3 bucket Optional: - [ ] ~~There is some sort of monitoring available to check Tamland executed successfully at the expected time~~ (not available in Dedicated) ## Status 2023-11-22 :white_check_mark: This is done - see https://gitlab.com/gitlab-com/gl-infra/scalability/-/issues/2622 for an example of how the data we generate looks like and can be displayed. The next step will be on the reporting side for Dedicated: https://gitlab.com/groups/gitlab-com/gl-infra/-/epics/1119