metrics(app): add app_block_gas_wanted and app_block_gas_wanted_ratio histograms (PLT-354)

[amir-deris]

Summary

Adds two OTel histogram instruments to make per-block gas utilisation observable — answering "how close are blocks getting to the MaxGasWanted cap in practice" without offline block parsing.

• app_block_gas_wanted (Int64Histogram, unit {gas}) — absolute totalGasWanted summed across all transactions in an accepted proposal
• app_block_gas_wanted_ratio (Float64Histogram, unit 1) — totalGasWanted / MaxGasWanted, stays meaningful if the cap is raised via governance

Both are emitted in checkTotalBlockGas (app/app.go) on the success path only — the function already short-circuits with return false when the sum exceeds the cap, so only complete, accepted totals are recorded. The existing app_failed_total_gas_wanted_check counter covers rejection events.

Bucket boundaries:

• Gas: 1 M–10 M in 1 M steps, then 5 M steps to 50 M (current cap)
• Ratio: 0.1 steps from 0.1 → 0.9, plus 0.95 and 1.0

Closes PLT-354.

Test plan

• Deploy to a test node and verify app_block_gas_wanted_bucket and app_block_gas_wanted_ratio_bucket appear in the Prometheus scrape. Can use this p50 query in Grafana by running
  histogram_quantile(0.5, rate({__name__="sei-chain_app_block_gas_wanted_ratio_bucket"}[10m]))
• Confirm values are non-zero and ratio stays ≤ 1.0 under normal load

[cursor]

PR Summary

Low Risk
Low risk: adds new OpenTelemetry histogram instruments and records them on the successful checkTotalBlockGas path, without changing acceptance/rejection logic.

Overview
Adds two new OTel histograms—app_block_gas_wanted (absolute total gas wanted per accepted proposal) and app_block_gas_wanted_ratio (utilisation vs MaxGasWanted)—with explicit bucket boundaries tuned for current gas caps.

Emits these metrics from checkTotalBlockGas in app/app.go only after the proposal passes gas-limit checks (and guards ratio recording against missing/zero consensus params).

^{Reviewed by Cursor Bugbot for commit 051930a. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
✅ passed	✅ passed	✅ passed	✅ passed	May 22, 2026, 7:56 AM

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.14%. Comparing base (d4fc9c5) to head (051930a).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3479      +/-   ##
==========================================
+ Coverage   59.11%   59.14%   +0.02%     
==========================================
  Files        2187     2186       -1     
  Lines      182237   181999     -238     
==========================================
- Hits       107730   107635      -95     
+ Misses      64851    64718     -133     
+ Partials     9656     9646      -10     

Flag	Coverage Δ
sei-chain-pr	50.58% <100.00%> (?)
sei-db	70.41% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
app/app.go	69.33% <100.00%> (-0.07%)	⬇️
app/metrics.go	15.38% <ø> (ø)

... and 43 files with indirect coverage changes

🚀 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.

[bdchatham]

nit: does this need to be a float?

[amir-deris]

Good point. I will convert to Int64Histogram for better semantics.

[bdchatham]

Seems like the PR description is slightly stale