Setting Priority Fees

Set dynamic priority fees using recent network percentiles and SetComputeUnitPrice so transactions compete during congestion.

Recipe

Quick-reference recipe card - copy-paste ready.

const fees = await rpc.getRecentPrioritizationFees([]).send();
fees.sort((a, b) => a.prioritizationFee - b.prioritizationFee);
const p75 = fees[Math.floor(fees.length * 0.75)]?.prioritizationFee ?? 1000;
 
// microLamports per CU in compute budget instruction
const microLamports = Math.min(p75 * 1.2, MAX_CAP);

When to reach for this:

• Mainnet launches and mints with contention.
• Relayer services batching user transactions.
• Wallets auto-tuning send success rate.
• Replacing static zero priority fee defaults.

Working Example

import { createSolanaRpc } from "@solana/kit";
 
const rpc = createSolanaRpc(process.env.RPC_URL!);
const MAX_MICRO_LAMPORTS = 500_000;
 
async function suggestPriorityFee(lockedAccounts: string[] = []) {
  const fees = await rpc
    .getRecentPrioritizationFees(
      lockedAccounts.map((a) => ({ account: a }))
    )
    .send();
 
  const samples = fees
    .map((f) => f.prioritizationFee)
    .filter((n) => n > 0)
    .sort((a, b) => a - b);
 
  if (samples.length === 0) return 1_000;
 
  const p75 = samples[Math.floor(samples.length * 0.75)];
  return Math.min(Math.ceil(p75 * 1.2), MAX_MICRO_LAMPORTS);
}
 
const microLamports = await suggestPriorityFee();
console.log("compute unit price:", microLamports);

What this demonstrates:

• getRecentPrioritizationFees samples landed transaction fees.
• 75th-90th percentile targeting balances cost vs success.
• Hard cap prevents runaway fees during fee spikes.

Deep Dive

How It Works

• Priority fee = microLamports * compute_units_requested (capped by limit instruction).
• Validators order transactions partly by total priority fee bid.
• Base signature fee (5000 lamports) remains separate from priority component.
• Provider APIs may return priorityFeeEstimate with percentile stats.

Strategy Tiers

TypeScript Notes

// Total priority lamports ≈ (microLamports * computeUnitLimit) / 1_000_000
// Add base fee separately in UX estimates

Gotchas

• Zero fee during hype mint - near-zero landing rate. Fix: dynamic percentile with floor.
• Uncapped p99 - users pay extreme fees. Fix: MAX_MICRO_LAMPORTS per product policy.
• Stale fee sample - market moved since fetch. Fix: refresh per send batch.
• High limit + high price - multiplies total priority lamports. Fix: right-size CU limit first.
• Devnet fee tuning meaningless - congestion unlike mainnet. Fix: test policy on mainnet small txs.

Alternatives

FAQs

Related

• Priority Fee & Simulation APIs - RPC data
• Priority Fees - conceptual model
• Jito Bundles - bundle tips
• Why Transactions Fail to Land - underpriced drops

Stack versions: This page was written for Agave 4.1.1, Solana CLI 3.0.10, Anchor 0.32.1, anchor-lang 0.32.1, Rust 1.91.1, @solana/kit 7.0.0, Surfpool 0.12.0, and LiteSVM 0.6.x.