feat: handle slashing state #51
Conversation
gusin13
requested review from
gitferry,
jrwbabylonlab and
jeremy-babylonlabs
| if dbErr != nil { | ||
| return false, types.NewError( | ||
| if dbErr := s.db.UpdateDelegationsStateByFinalityProvider( | ||
| ctx, fpBTCPKHex, types.StateSlashed, |
There was a problem hiding this comment.
what if service restart in the mid of this process? e.g there are 1m delegations under the FP PK. we only processed 10k of those then we deployed the service again
There was a problem hiding this comment.
hmm so if this happens then last processed height not updated.
in next restart, it picks up from same height (at which it failed in prev run), finds this slashing event and then tries to make a bulk update again 🤔
There was a problem hiding this comment.
ah, but slashing operation is very time consuming. we shall not couple it with the main processing thread.
Given the same example above, say u have 1 million delegations under this FP. our indexer will be blocked on this operation until all delegations are updated.
My proposal is to perform this slashing operation in a different go routine, you can do this by either:
- Emit "slashing msg" into rabbitmq, then process it by indexer itself. only delete the msg once all delegations are processed.
- OR, you create a temporary db collection just for slashing. let's say
slashing_registry. then a go routing monitoring this collection and remove the entry once no more delegations to process.
There was a problem hiding this comment.
i see, sure created a ticket to optimize the slashing process
#57
Intercepts the slashing event from Babylon and then transitions all delegations for the slashed fp to "Slashed" state.
Note -
Currently Babylon doesn't emit slashing tx so indexer cannot identify when the slashing output is spent and the slashing withdrawal has happened or not.
Have added a todo for now regarding this, also created a ticket on babylon
babylonlabs-io/babylon#271