Current behavior

The node is accepting load and responding with 200 status on /status endpoint even though indexes are not up to date.

Expected behavior

The node will only accept traffic and respond with status 200 after the indexes are up to date.

Notes

Before 9.1.1 the node was catching up with the changes after node startup in asynchronous manner using replicatedindexoperation table and NodeReindeService.

The status endpoint was not indicating that the indexes are not 100% up to date and was allowing load to the node by replying 200 status code and state: "Running"

The idea is to not let startup of the node proceed until the node has a full index available.

Right now it's common to enforce such behavior by other means outside of Jira (for example by monitoring /index/summary ednpoint)