Handle large chain reorgs with >32 slots about dora HOT 3 CLOSED

With 3494311, epoch aggregations will now be re-processed automatically when a large reorg happens and the relevant epochs are still available in memory (inMemoryEpochs setting).

If large reorgs are expected, set the inMemoryEpochs setting to a higher value.

from dora.

I'm about to completely re-implement the indexer related code (forkmon branch).
The new code works quite different and allows defining multiple endpoints.
These endpoints are used in parallel to keep track of the chain and possible chain forks.

Epoch processing & orphaned detection are now done for finalized epochs only.
Non-finalized slots are maintained in memory (with block bodies being saved to the db to save memory if non-finality gets too long)
This approach looks much more reliable so far

from dora.

#8 should fix any reorg related problems.
The explorer should now be able to handle any kind of reorg properly.

from dora.