Argument-0008: Retention at Rank 2


Report Date	Date of submission (2026/05/12)
Submitted by	Tsvetomir Dimitrov

Member details

Matrix username: @tsvetomir:parity.io
Polkadot address: 1QhVP5qzR2LfXqP77N1JcuwHoY7NH8JVRNFm1hSooE9d4pR
Current rank: 2
Date of initial induction: 2023/10/11
Date of last report: 2026/02/25
Area(s) of Expertise/Interest: Disputes, Collator protocol

Reporting period

Start date: 2025/12/08
End date: 2026/02/25

Argument

During this reporting period, my primary focus was still the collator protocol revamp and the malus-collator (testing tool for the revamp).

I have fixed a few collator revamp related issues discovered during audits and testing:

Each validator has got a limited number of peer scores it can track. When this limit is reached some peers need to be pruned from the database in order to keep it in a reasonable size. The initial implementation of pruning gave edge to new peers because it used only the timestamp of the last bump when evicting peers to the DB. As a result, a high score collator which was inactive for a while can easily be evicted by new peers with minimal score. To fix this a the ratio of calculate score / time_since_last_bump is calculated for each peer and the one with minimal score is evicted from the DB. The change is implemented in polkadot-sdk pr-11576.
When starting the node with a warp sync and we hit a period near the WARP_SYNC_TARGET_BLOCK (each 512 blocks) we might not be able to call a runtime apis for some blocks, which will yield an error in the collator protocol revamp. To prevent the subsystem from exiting these errors aren't bubbled up anymore. Implemented in polkadot-sdk pr-11496.
Decrease the log level of some noisy logs - polkadot-sdk pr-11417.
Sync the values of CONNECTED_PEERS_PARA_LIMIT in experimental collator protocol and MAX_AUTHORITY_INCOMING_STREAMS in peer_set. Implemented in polkadot-sdk pr-11762.
Simpler advertisement fetching mechanism. polkadot-sdk pr-11008

In malus-collator the most notable improvement is the implementation of the automatic targets discovery. Initially the malus-collator required a list of peer ids of its target. With this change it supports automatic discovery of the validators of a given para plus tracking them on session changes. I can't provide a link to the PR since the project is not open sourced yet.

Voting record

Ranks	Activity thresholds	Agreement thresholds	Member's voting activities	Comments
I	90%	N/A
II	80%	N/A	I have voted on 0 out of 0 referenda in which I was eligible to vote (i.e 100 % voting activity). Out of 0 referenda in which members of higher ranks were in complete agreement, I have voted in line with the consensus 0 times (i.e 100 % voting agreement).	There were no referendas I could participate in during the reporting period
III	70%	100%
IV	60%	90%
V	50%	80%
VI	40%	70%

Misc

Question(s):
Concern(s):
Comment(s):