db: add composite indexes for telemetry query performance

[codewithzubair07]

Closes #347

Problem

Analytical dashboard queries bog down when scanning unindexed columns
in large telemetry tables, especially for deep historical metrics.

Solution

Added composite index migration files in src/database/migrations/:

Files added

• 001_telemetry_query_indexes.sql — 10 composite indexes across 6 tables
• 001_telemetry_query_indexes_rollback.sql — rollback script

Indexes added

Table	Index	Query accelerated
PriceHistory	source, timestamp	Provider history over time
PriceHistory	currency, source, timestamp	Dashboard per-pair per-provider
OnChainPrice	currency, ledgerSeq	Ledger history pagination
OnChainPrice	currency, confirmedAt	Time-range dashboard queries
RawData	provider, currency, fetchedAt	Validator/provider telemetry
OhlcCandle	currency, granularity, openTime + INCLUDE	Sparkline charts
ErrorLog	providerName, occurredAt	Recent errors per validator
ProviderReputation	status, reliabilityScore	Validator ranking
ProviderReputation	providerName, lastUpdated	Staleness monitoring
ComplianceMetadata	eventType, resolved, createdAt	Latency violation dashboard

All indexes use CONCURRENTLY to avoid table locks in production.