#Forgive me - as a newer user, I am only allowed to place two links in my posts.
#I had many more references in my original write up.
A somewhat popular use-case for Elasticsearch, and other time-series data stores, is long term metric storage of metrics gathered by Prometheus.
See “Prometheus Integrations” for some examples: https://prometheus.io/docs/operating/integrations/
In particular, Elastic supports gathering Prometheus-gathered metrics via the remote_write feature in metricbeat: https://www.elastic.co/guide/en/beats/metricbeat/current/metricbeat-metricset-prometheus-remote_write.html
That said - many of the aforementioned solution patterns do not support remote_read, nor do they have a way to continue to query the data using PromQL, the Prometheus querying language.
I would like to request three features to promote OpenDistro as a premiere datastore for Prometheus collected metrics.
Add native support for PromQL and the querying api.
This feature could likely be largely cloned from Prometheus source code - which is licensed under the Apache License 2.0 license. There would likely need to be adjustments to accommodate index schemas, security features, and obviously using Elasticsearch as the datastore instead of the normal Prometheus database.
As part of the Feature 1, you would likely need to create a Prometheus remote_read adapter.
It would be beneficial to the community if this adapter were made more generic, and separately available, thus allowing generic Prometheus to use OpenDistro via the remote_read adapter.
This should help better promote OpenDistro among the Prometheus community and should help drive adoption of OpenDistro as a long term metric store.
While there are a number of ways data could get pipelined to OpenDistro from the Prometheus ecosystem (eg. the metricbeat method described above), to really push OpenDistro ahead, it should directly support ingestion of metrics via Prometheus’s remote_write mechanism.