Disaster recovery for Loki

giantswarm / roadmap

Giant Swarm Product Roadmap

Apache License 2.0

3 stars 0 forks source link

Open QuentinBisson opened 7 months ago

QuentinBisson commented 7 months ago

name: Disaster recovery for Loki about: This Epic is about support disaster recovery mechanism for Loki

- As a user, I want to be able to access/store my logs even if the management cluster is down

Please give all possible details, requirements, etc. Technical or not.

Please link issues/epics here which have to be solved before

hervenicol commented 7 months ago

Ideas:

if a new MC is created, use it to access old logs. We could have a schema_config pointing to the old S3 for data older than old WC collapse/new WC creation. :warning: it requires IRSA to allow access to old MC's S3.
we could also deploy a "read-path-only" loki anywhere, that would read data from old S3. Requires:
- access to S3 for the old MC (IRSA?)
- a grafana datasource