I'm not sure it makes sense to add tenant- to each of the error IDs. The limits don't necessarily correspond to user-specific limits, they can also come from the limits section of the global configuration file: https://grafana.com/docs/mimir/latest/operators-guide/configuring/reference-configuration-parameters/#limits

I'm not sure it makes sense to add tenant- to each of the error IDs. The limits don't necessarily correspond to user-specific limits, they can also come from the limits section of the global configuration file: https://grafana.com/docs/mimir/latest/operators-guide/configuring/reference-configuration-parameters/#limits

The tenant- prefix originated by a feedback I gave to Jeanette. After looking these changes, I want to step back. I agree with Nick that looks "too much". I think the only one where we need to clarify it is TooManyHAClusters, so I would add the tenant- prefix just to that. I think other errors have no need to disambiguate between per-tenant and per-instance (or per-something-else).