mempool to slasher stats bridge

[Mododo]

Model (pre-review): anchor stats belong to blocks in a slasher batch.
Corner cases:

• there is a received_and_collated=true fast path, when block is both collated and received from network. Then validation step is omitted and signatures are not reported to slasher. We actually have mempool stats for that block, but currently they are not reported.
• in case block collation was cancelled by a received block (validate_block() returns ValidationStatus::Skipped) anchor stats are collected but not reported to slasher too.

Slasher contract is left untouched intentionally until we agree on the models and code flow

The drawbacks of suggested design are:

• at every subset change mempool has to omit max_consensus_lag_rounds=210 rounds of anchor stats, otherwise old vset validator_idxs will be mixed with new one (mempool makes no difference between vset and subset changes), so subsets should not be changed too often
• stats attributed to anchor are older and don't belong to that anchor history: stats are for dag rounds GC'ed after that anchor and have at least commit_history_rounds=20 rounds lag

Inherited from general design: stats may not be reported to slasher near validation session end because detached validation tasks can finish after the session has already been closed on the slasher side; same for already flushed batches during the same session. Also in case a collated block is replaced by accepted by the quorum, its stats remain in slasher.

Parent branch code suggestion #1068: in ValidatorSessionScope.remap_ids use PeerId to slot map (instead of validator_idx->slot) to allow cross-subset stat collection in mempool (like references_skipped: see commit fix(consensus): mempool stats across subset change)

---

Updates after review

Model: anchor stats are imported during a slasher batch (disregarding block collation).

Old corner cases are removed.
New corner cases:

• collector batch was advanced only by validation progress (on_block_validated() or on_block_skipped()) - batches didn't overlap, so anchors imported for the first block were ignored
  • see feat(slasher): batches overlap; shouldn't we convert it into a PR to parent branch first, prior TODO: Split or grow the previous batch to not discard events?
• stats of anchors imported by try_import_init_anchors() are ignored
• anchors_cache.remove_last_imported_above(next_chain_time) removes anchors from block input, but stats are already reported - are anchors imported the same way on every node?

[github-actions]

🧪 Network Tests

To run network tests for this PR, use:

gh workflow run network-tests.yml -f pr_number=1067

Available test options:

• Run all tests: gh workflow run network-tests.yml -f pr_number=1067
• Run specific test: gh workflow run network-tests.yml -f pr_number=1067 -f test_selection=ping-pong

Test types: destroyable, ping-pong, one-to-many-internal-messages, fq-deploy, nft-index, persistent-sync

Results will be posted as workflow runs in the Actions tab.

[codecov]

Codecov Report

❌ Patch coverage is 28.14570% with 217 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (feature/slasher-session@587cd70). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
slasher/src/collector/validator_events.rs	0.00%	64 Missing ⚠️
slasher/src/storage/models.rs	0.00%	35 Missing ⚠️
slasher/src/bc/stub_contract.rs	29.16%	30 Missing and 4 partials ⚠️
consensus/src/engine/committer_task.rs	0.00%	26 Missing ⚠️
slasher-traits/src/validator.rs	0.00%	17 Missing ⚠️
slasher/src/bc/mod.rs	48.48%	17 Missing ⚠️
collator/src/collator/mod.rs	12.50%	14 Missing ⚠️
collator/src/collator/do_collate/finalize.rs	89.58%	5 Missing ⚠️
cli/src/node/mod.rs	0.00%	1 Missing ⚠️
collator/src/mempool/impls/common/shuttle.rs	0.00%	1 Missing ⚠️
... and 3 more

Additional details and impacted files

@@                    Coverage Diff                     @@
##             feature/slasher-session    #1067   +/-   ##
==========================================================
  Coverage                           ?   58.06%           
==========================================================
  Files                              ?      465           
  Lines                              ?    78608           
  Branches                           ?    78608           
==========================================================
  Hits                               ?    45647           
  Misses                             ?    30838           
  Partials                           ?     2123           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[SmaGMan]

anchor_stats_history is sorted here, but signatures_history is not sorted. But seems like it is required for them to be with the same sort order in the TLRead::read_from implementation

[Mododo]

That's right, because mempool emits stats already sorted and they are just passed as-is through collator.
Mempool stats don't need additional layer of validator_idx|PeerId remapping and sorting - there's a PeerSchedule inside mempool to keep all v(sub)set data

[Mododo]

misunderstood at first, fixed