This message was deleted.

My experience with metrics in Opensearch/Elasticsearch has been that they are OK, but in terms of storage efficiency and performance they're not as good as a dedicated TSDB, so yeah I think Cortex will scale better. TBH Prometheus based solutions like Cortex are probably not the best TSDB either. Post GA one of the things we will be doing is reviewing alternatives there. In terms of alerting I think Alertmanager is the gold standard there. However with the alerting for logging that we are working on we will be building on the built in alerting.