KVORUM
lidoSource: Snapshot · #0x969f9a3b3f52742b9763d8e128f2418a0ce888433e130a60f7fef809787ffdb4

Update Lido on Ethereum Standard Node Operator Protocol - Validator Exits

succeededDec 23, 2024, 04:00 PM UTCsignaling · non-bindingbyDB0xdbbc…1cb6View on Snapshot ↗

01Summary

AISummary by Kvorum

An AI summary is not generated yet. Read the description →

02Description

1,497 words

Summary

This proposal seeks the community’s approval to update the currently in effect Lido on Ethereum (LoE) Validator Exits Policy 1.0 (Snapshot) to the proposed Standard Node Operator Protocol (SNOP) on Validator Exits 2.0 (diff-view), following the draft as presented and discussed on the Lido Research Forum and during two community town hall calls.

Context and Motivation

The current Validator Exits Policy is outdated as it does not reflect protocol improvements made in Staking Router (SR) 2.0 (Snapshot, on-chain vote), and the addition of the Simple DVT (Snapshot, on-chain vote) and Community Staking (Snapshot, on-chain vote) modules which are now live on mainnet.

SR 2.0 includes some new concepts and individual approaches with regard to validator exits with respect to the now three active Staking Modules (SMs): the Curated Module (CM), the Simple DVT Module (SDVTM), and the Community Staking Module (CSM).

The Validator Exits SNOP has been reworked with a logical structure that can be easily expanded, to coherently describe the key concepts and mechanisms around validator exits, both in Ethereum and specifically to Lido, to define Node Operator (NO) expectations with regard to the usage of the protocol, and the possible consequences of non-conformance as per the rules of each SM.

Overview of Changes and Additions

The changes and additions include:

  • Refactoring of the structure in an attempt to create a template structure for all future SNOPs, and re-organization of some ancillary content to appendices
  • Extension of the scope to NOs using the Simple DVT Module and/or Community Staking Module
  • Definition of the terminology specific to SDVTM and CSM
  • Introduction of the validator exit order of SR 2.0
  • Introduction of SM stake share limits and explanation of their effect on exit priority
  • Introduction of NO target limits and explanation of their effect on exit priority
  • A clearer description of a mechanism through which an NO validator exit request status of delinquent can recover to an in good standing status
  • A clearer explanation, on a per-SM basis, of validator exit requirements and penalties for non-conformance as per the protocol code

Key Excerpts

B - Scope

This SNOP applies to LoE, the NOs who use one or more of the Staking Router (SR)'s Curated Module (CM), Simple Distributed Validator Technology Module (SDVTM), and Community Staking Module (CSM) and the validators the NOs operate as part of the Liquid Staking Protocol (LSP).

[…]

C - Validator Exit Order

[…]

In the SR 2.0, the priority of exits is determined by the following sorting predicate.

SortingStaking ModuleNode OperatorValidator
VLowest number of delayed validators
VHighest number of targeted validators to boosted exit
VHighest number of targeted validators to smooth exit
VHighest deviation from the exit share limit (i.e. highest positive difference between currentShare and priorityExitShareThreshold)
VHighest stake weight
VHighest number of active validators
VLowest index

[…]

D - Node Operator Responsibilities

LoE contains on-chain signaling mechanisms that serve to notify NOs when to process validator exits. If exits for keys under their operation have been requested through the Validators Exit Bus Oracle (VEBO) smart contract, NOs have the duty to fully, correctly, and timely exit validators as determined by the requirements and rules set by and for any of the reasons described in the sections A - Purpose and C - Validator Exit Order, as per the latest in force SNOP.

To determine whether NOs have appropriately executed the required actions, this SNOP outlines the operative timelines for exits processing, and specifies the consequences of non-conformance. Generally, NOs are expected to exit the indicated validators in a reasonable time frame. The actual mechanisms for validators to be exited are at the discretion of the NOs, with community-built open-source tooling (see Appendix B) to aid the NOs in identifying and processing signaled requests being available in addition to any internal tooling, APIs, or manual processes NOs may use.

D.1 - Out of Order Exits

Out of order exits refer to exits that are made by an NO of one of or both the CM and SDVTM when a validator exit request has not been made through the VEBO. In such cases, the NO must notify the community via the Lido Research Forum that an out of order exit has been processed, how many and which validators have been exited, and the reason for it. As the CSM is permissionless and validators are bonded, this expectation is not extended to validators contained therein, CSM NOs are free to exit out of order at their discretion.

D.2 - Business Continuity

If, at any time, an NO of one of or both the CM and SDVTM is unable to continue operations, e.g. has become insolvent, the NO must notify the community via the Lido Research Forum of the circumstances and signal intent to exit all of the validators the NO operates using LoE. If LDO token holders, via a ratified vote within 8 days, do not instruct the NO otherwise, the NO must proceed with triggering all of the exits.

D.3 - Validator Exits Performance

[…]

With respect to validator exits performance, at any point in time, each NO may be considered to have one of the below three validator exit request statuses:

  • in good standing – validator exit requests are being processed fully, correctly, and timely
  • delayed – validator exit requests are being processed incompletely, incorrectly, or not within the desired time frame
  • delinquent – validator exit requests are being processed incompletely, incorrectly, or not within the acceptable time frame

For every delinquent exit request, i.e. a request that has been unprocessed for longer than the delinquency threshold, an NO's per-SM stuckKeysCount increases by 1. An NO with a stuckKeysCount > 0 is delinquent, which may result in the NO's share of rewards being reduced and/or penalties imposed for the affected and possibly further frames, as determined by the respective SM. The time for an NO to recover to in good standing depends on four aspects, (1) the NO having processed all delinquent exit requests, (2) the frame length of the Lido Oracle sub-module that reports updates on validator exit request statuses, (3) the frame length of the sub-module whose reports account for NO penalties related to unprocessed exit requests or delinquency, and (4) any SM-specific cooldown period for NO delinquency. For the CM and SDVTM, the relevant sub-module is the Accounting Oracle (AO), for the CSM it is the Performance Oracle (PO). The CM and SDVTM have a cooldown period for NO delinquency of 120 hours, long enough to determine whether, immediately after service restoration, subsequently received validator exit requests are processed timely. The CSM does not have such a cooldown period.

To avoid NO delinquency, NOs must adhere to the following time frames.

EventRequirement to not be considered delayedRequirement to not be considered delinquent
Processing of signalled validator exit requestsAll signalled requests are processed in less than 24 hours after a VEBO reportSome signalled requests are taking longer than 24 but less than 96 hours to process
Escalation of inability to process signalled validator exit request with reasonASAP but no longer than 24 hoursASAP but no longer than 96 hours

E - Consequences of Non-conformance

In case that an NO is not processing validator exit requests timely, the below actions shall be taken. Separate measures may be applied depending on the technical characteristics of the SM within which the request originated. For example, while NOs are known for the permissioned CM and SDVTM, NOs of the permissionless CSM may remain anonymous and reaching the NOs is only possible if a contact is voluntarily shared with the community.

ConsequencesNonconformance.png

[…]

The full text of the proposed Validator Exits SNOP 2.0 published to the IPFS will be a single source of truth if the proposal passes. The alternative sources that can be used to access the policy text are as follows:

Vote Outcomes

If the "For" option reaches quorum and a majority in favor of the proposal, the new version published to the IPFS will replace the previous document one-to-one and be used as the human-legible reference for the interaction of the protocol with NOs using LoE to operate validators from the moment of the conclusion of the vote.

If a quorum is not met or the proposal is rejected, 1.0 remains in effect until a quorum is reached in a subsequent vote and/or the draft is adjusted to satisfy the community.

03Mismatch analysis

AIMismatch analysis by Kvorum

A calldata-vs-description analysis is not generated yet. Review the decoded actions →

04Actions

0 total

This proposal carries no on-chain actions.

05Tally

Summed from votes

Voters120
VP participating0
Outcome (current)For
Current support100%

06Voters

120 total
Filter:
VoterChoice% of totalRationale
4A0x4af8…6a0bFor023.21%—2y ago
810x8172…922eFor010.98%—2y ago
7E0x7ee0…bb6cFor09.02%—2y ago
LIlido.banteg.eth0x55bc…f0faFor06.47%—2y ago
6D0x6d8d…0cccFor06.39%—2y ago
A40xa418…b737For06.35%—2y ago
510x510a…8d0eFor06.17%—2y ago
B90xb933…8dd1For04.81%—2y ago
3D0x3ddc…05b0For04.58%—2y ago
FF0xff41…a67bFor04.49%—2y ago
CE0xce3b…2e55For03.59%—2y ago
BE0xbeb3…7646For03.42%—2y ago
6A0x6a39…4bd2For02.39%—2y ago
1F0x1f76…5ed6For02.01%—2y ago
7A0x7a59…8e22For01.24%—2y ago
A20xa2ff…7be9For01.12%—2y ago
B60xb664…f890For00.62%—2y ago
870x8787…ea52For00.58%—2y ago
E80xe80e…28e8For00.57%—2y ago
2C0x2cc1…4df1For00.50%—2y ago
6F0x6f9b…8a1dFor00.36%—2y ago
100x10a2…36f3For00.25%—2y ago
620x624d…2cd0For00.22%—2y ago
980x9830…bd94For00.14%—2y ago
060x06a9…bba0For00.14%—2y ago
3D0x3d63…4afcFor00.11%—2y ago
780x78d1…753aFor00.07%—2y ago
610x61cc…87d1For00.05%—2y ago
5F0x5fdc…b69fFor00.04%—2y ago
3F0x3fb1…4c8aFor00.02%—2y ago
500x50ed…e1b2For00.02%—2y ago
9E0x9e0b…da3aFor00.01%—2y ago
050x0579…4048For00.00%—2y ago
420x42e6…3fb0For00.00%—2y ago
8A0x8ab6…e0baFor00.00%—2y ago
7A0x7af3…8df3For00.00%—2y ago
4D0x4d26…7017For00.00%—2y ago
DB0xdbbc…1cb6For00.00%—2y ago
3D0x3d5a…b97fFor00.00%—2y ago
5E0x5e45…3248For00.00%—2y ago
6A0x6a2d…03e0For00.00%—2y ago
580x5820…a011For00.00%—2y ago
2F0x2f15…631fFor00.00%—2y ago
5A0x5a51…347cFor00.00%—2y ago
B70xb792…89d7For00.00%—2y ago
ED0xed95…68a6For00.00%—2y ago
520x5243…3440Against00.00%—2y ago
EC0xecc2…6ac9For00.00%—2y ago
E60xe6b5…f1fdFor00.00%—2y ago
C10xc1c2…24f8For00.00%—2y ago
Showing 50

07Forum

No forum thread is confidently linked to this proposal.

08Similar proposals

AI · embeddings

No semantically similar proposals found yet.