Elastic Search pools for REST API require duplicated, expensive, calculations

Where?

REST API Public, Polite, Plus, Internal pools.

What's the situation?

Each pool is effectively firewalled with respect to the other pools. This means that every process (indexing, coverage indexing) needs to run in parallel. This is duplicate work that could theoretically be done once.

What does it make more difficult?

Coverage data might drift, and might be more expensive.

How can we improve it?

• Calculate the coverage data upstream and distribute out to Elastic Search clusters.