Loading...

XML

Word

Printable

Feature description:

given:
- I have several Optimize instances running with the same version which use the same Elasticsearch instance
- Optimize instance A is used to import the data and offers the user interface to create, edit and delete entities (e.g. reports or dashboards)
- Optimize instance B is configured to only offer the user interface and the import is disabled.
when:
- Optimize instance A is failing or shutdown (e.g. due to maintenance work)
then:
- I can still use B to create, edit, update or delete entities in Optimize
- The reports are all based on the current state of the data without new data being imported to Optimize
such that:
- The user has the possibility to always ensure a version of Optimize up and running, so in case one instance fails, the other one can take over.

AT:

optimize can be run clustered with two or more instances
the cluster setup specifics are documented (dedicated guide + new config params)

Hint:

Make sure that there are no conflicts during the index creation while running multiple instances in parallel
When the import is disabled the status api should return that Optimize is currently not importing
It should be documented how the clustering works, what are the use-cases and limitation of it

depends on

OPT-1918 Implement a distributed user session

1.	Disable import by engine in config	Done	Unassigned
2.	Add Optimize Clustering test pipeline	Done	Unassigned
3.	Clustering section in Technical Guide	Done	Unassigned

I can setup a failover scenario for Optimize