Pick a recent mainnet slot (within the last 7 days). Document everything about that slot's block: the leader, the number of transactions, the compute units consumed, the priority fees collected, and a breakdown of the top 5 programs invoked by CU usage. Produce a one-page memo.
Use @solana/web3.js getBlock with maxSupportedTransactionVersion: 0 to handle v0 versioned txs.
For program CU breakdown: inspect each tx's meta.logMessages for Program <ID> consumed <N> of <M> compute units lines — that's the per-program CU.
Solscan's "Top Programs" widget is a fine sanity check.
Slots near "token launch hours" (US morning / EU afternoon) are noisiest. Try a slot ~3am UTC for clearer signal.
Slot: 263456789 | Block time: 2026-05-13 14:22:11 UTC
Leader: GgreL...XXp (stake: 2.3M SOL, 0.4% of total)
Tx count: 1,247 | CU consumed: 38.2M / 48M limit (79.5%)
Total fees: 0.412 SOL (base: 0.006 + priority: 0.406)
Top programs by CU:
1. Jupiter Aggregator V6 8.4M CU (22%)
2. Raydium AMM v4 5.1M CU (13%)
3. Orca Whirlpool 3.2M CU (8%)
4. Pyth Oracle 2.8M CU (7%)
5. Token Program 2.1M CU (6%)
Memo: Slot was dominated by DEX activity; 50% of CU went to AMMs.
Pyth's high CU suggests an active price-update round.