fix: overfilled blocks should report max fullness instead of 0

[mpskowron]

Overview

Currently, if a block gets overfilled, its fullness is reported as 0% instead of 100%. This happens because SyntheticCost::normalize() returns None when any dimension exceeds limits, and the code falls back to NormalizedCost::ZERO.

This incorrect reporting affects fee adjustment calculations - the algorithm would treat an overfilled block as empty, leading to lower fees when they should be higher due to congestion.

This PR clamps costs to limits before normalizing, ensuring overfilled blocks report maximum fullness.

⚠️ Toolkit Backwards Compatibility

When the toolkit replays historical blocks to build its local ledger state:

Blocks	On-chain state	Toolkit replay
Before upgrade	Old logic (0% for overfull)	New logic (100%)
After upgrade	New logic (100%)	New logic (100%)

If overfilled blocks exist in chain history, the toolkit's local state will diverge from on-chain for those blocks.

Questions:

• Do overfilled blocks exist on any network?
• If they do, should we add a version gate (use old logic before upgrade block)?
• how was it resolved last time? (I see that it was change already on 5th of January)

🗹 TODO before merging

• Ready
• Resolve toolkit backwards compatibility approach

📌 Submission Checklist

• Changes are backward-compatible (or flagged if breaking)
• Pull request description explains why the change is needed
• Self-reviewed the diff
• I have included a change file, or skipped for this reason:
• If the changes introduce a new feature, I have bumped the node minor version
• Update documentation (if relevant)
• Updated AGENTS.md if build commands, architecture, or workflows changed
• No new todos introduced

🧪 Testing Evidence

Please describe any additional testing aside from CI:
Local load test sending over 20 shielded token mint transactions to fill the block (had to modify the code to allow block overfilling). The error message was displayed with block clamped correctly to the limits.

• Additional tests are provided (if possible)

🔱 Fork Strategy

• Node Runtime Update
• Node Client Update
• Other: See the Toolkit Backwards Compatibility point above
• N/A

Links

Ticket: https://shielded.atlassian.net/browse/PM-20839

[github-actions]

KICS version: v2.1.16

Category Results CRITICAL 0 HIGH 0 MEDIUM 94 LOW 12 INFO 83 TRACE 0 TOTAL 189	Metric Values Files scanned 28 Files parsed 28 Files failed to scan 0 Total executed queries 73 Queries failed to execute 0 Execution time 6

[ozgb]

Missed the concerns raised in the description r.e. the toolkit - we'll need to update it so that the logic matches

EDIT: I misunderstood - the issue was instead that chains with existing overfilled blocks would cause the toolkit to fail to sync. @mpskowron is checking this

[mpskowron]

Missed the concerns raised in the description r.e. the toolkit - we'll need to update it so that the logic matches

EDIT: I misunderstood - the issue was instead that chains with existing overfilled blocks would cause the toolkit to fail to sync. @mpskowron is checking this

@ozgb I've done the test, added details to the ticket: https://shielded.atlassian.net/browse/PM-20839?focusedCommentId=56688

[ozgb]

LGTM