Proposal Fixing The Last Slashing Event On Kusama

What happened?

• On 9.12.2022, a large amount of disputes flooded the network.
• Other validators had issues handling the flood of messages, the disputes-coordinator became unresponsive, causing other subsystems to shut down. The disputes-coordinator task never fully finished, preventing the node to completely shutdown (where it might usually be automatically restarted depending on the setup). While in a zombie state, the node wasn't able to produce imonline heartbeats or produce blocks unless the validator manually restarted the node, causing many offline events.
• Enough validators had offline events for there to be correlated slashing.
• As a consequence of the amount of validators offline, election results also failed, whereby the only way for elections to happen again is for an election solution to be submitted via governance.
• Elections can only start again with governance is restarted.
• Chilled validators wont be able to validate again until the election via governance is restarted.

What is this proposal meant to execute?

From the validator set PoV there are 3 action items to execute:

1. Cancelling the slash so that no account balances get changed.
2. Resuming elections, which will work but with a validator set that is relatively weak. This brings us to point 3:
3. Reversing the chilling, which is unfortunately difficult to do given how the data is laid out on-chain (we need to change one filed in a struct in a map, and doing it via system::set_storage is tricky because there could be more slashes in the meantime). A PR is ready to execute this, but a new runtime will be needed for this last point.

This proposal aims to execute 3 calls :

• canceling the slash
• lowering min_untrusted_score by 20% (to help deal with the low quality elections in the meantime)
• force set the validator set, represented by the full exposures

Proposal Fixing The Last Slashing Event On Kusama

What happened?

• On 9.12.2022, a large amount of disputes flooded the network.
• Other validators had issues handling the flood of messages, the disputes-coordinator became unresponsive, causing other subsystems to shut down. The disputes-coordinator task never fully finished, preventing the node to completely shutdown (where it might usually be automatically restarted depending on the setup). While in a zombie state, the node wasn't able to produce imonline heartbeats or produce blocks unless the validator manually restarted the node, causing many offline events.
• Enough validators had offline events for there to be correlated slashing.
• As a consequence of the amount of validators offline, election results also failed, whereby the only way for elections to happen again is for an election solution to be submitted via governance.
• Elections can only start again with governance is restarted.
• Chilled validators wont be able to validate again until the election via governance is restarted.

What is this proposal meant to execute?

From the validator set PoV there are 3 action items to execute:

1. Cancelling the slash so that no account balances get changed.
2. Resuming elections, which will work but with a validator set that is relatively weak. This brings us to point 3:
3. Reversing the chilling, which is unfortunately difficult to do given how the data is laid out on-chain (we need to change one filed in a struct in a map, and doing it via system::set_storage is tricky because there could be more slashes in the meantime). A PR is ready to execute this, but a new runtime will be needed for this last point.

This proposal aims to execute 3 calls :

• canceling the slash
• lowering min_untrusted_score by 20% (to help deal with the low quality elections in the meantime)
• force set the validator set, represented by the full exposures

For the sake of posterity, this is the branch and the code snippet that I used to generate the proposal. Sorry for the delay:

https://github.com/paritytech/polkadot/compare/release-v0.9.34...kiz-fix-kusama-slash?expand=1