Fix Total Stake hero value to use validator staking pool

[ignacio-rsg]

Summary

Fixes the dashboard hero Total Stake metric on Lava Info.

Previously, /index/stats calculated totalStake by summing provider stake plus provider delegation across all providers:

provider.totalStake + provider.totalDelegation

That made the dashboard display a provider restake/delegation total instead of the validator staking total expected by the UI label.

This PR changes totalStake to use the official Cosmos staking pool value from:

/cosmos/staking/v1beta1/pool

Specifically, the dashboard now uses:

pool.bonded_tokens

This matches the ticket expectation of approximately 381M LAVA.

What Changed

• Updated GET /index/stats to fetch the staking pool via the existing fetchStakingPool() helper.
• Changed totalStake to return stakingPool.bonded_tokens.
• Kept fetchAllProviders() in place for activeProviderCount.
• Updated the /index/stats route test to ensure totalStake comes from the staking pool, not provider stake/delegation totals.

What Did Not Change

The rest of the dashboard stat logic is unchanged:

• totalCu still comes from Indexer GraphQL all-time relay aggregates.
• totalRelays still comes from Indexer GraphQL all-time relay aggregates.
• cu30d still comes from Indexer GraphQL 30-day aggregates.
• relays30d still comes from Indexer GraphQL 30-day aggregates.
• activeProviderCount still comes from fetchAllProviders().
• latestBlock still comes from fetchLatestBlockHeight().

Local Verification

Verified locally that /index/stats now returns:

"totalStake": "381345490217757"

which renders as approximately:

381.35M LAVA

Verified source endpoint:

https://lava.rest.lava.build/cosmos/staking/v1beta1/pool

returns:

{
  "pool": {
    "not_bonded_tokens": "41395102178952",
    "bonded_tokens": "381345490217757"
  }
}

Tests

Passed:

pnpm --filter @info/api exec vitest run src/__tests__/routes-index.test.ts
pnpm --filter @info/api test
pnpm --filter @info/api typecheck
pnpm --filter @info/web typecheck

Notes

The local dashboard may show 0 for CU and relay metrics unless a real INDEXER_GRAPHQL_URL is configured. That is unrelated to this fix. Those values still use the same existing indexer logic.