Question 1

What causes blockchain congestion?

Accepted Answer

TL;DR: Blockchain congestion occurs when more transactions are submitted to the network than can fit in the next block. Every blockchain has a maximum amount of computation or data it can process per block, and when demand exceeds that limit, a backlog forms in the mempool (the waiting area for unconfirmed transactions). Congestion leads to higher transaction fees (as users compete for limited block space), longer confirmation times, and degraded user experience. The root causes include fixed block capacity, demand spikes from popular events, fee market dynamics, and the fundamental design tradeoffs that prioritize security and decentralization over raw throughput. The Simple Explanation A blockchain processes transactions in batches called blocks. Each block has a capacity limit. On Ethereum, this limit is expressed as a gas limit (currently around 30 million gas per block). On Bitcoin, it is a weight limit (4 million weight units per block). On Solana, it is a combination of compute units and account locks per slot. Regardless of the specific mechanism, every chain has a ceiling on how much work it can do per block. When the number of transactions being submitted to the network is lower than the block capacity, everything runs smoothly. Transactions are included in the next block, fees are low, and confirmation is fast. When the number of transactions exceeds the block capacity, a queue forms. This queue is the mempool, a holding area maintained by each node where valid but unconfirmed transactions wait for inclusion in a future block. The mempool is not first-come, first-served. It is a fee-based priority queue. Block producers (miners or validators) select the transactions that pay the highest fees because that maximizes their revenue. When the mempool is full, users who want their transactions confirmed quickly must outbid everyone else by offering a higher fee. This bidding war is what causes gas prices to spike during congestion. Users who are unwilling or unable to pay the elevated fees have their transactions stuck in the mempool for minutes, hours, or sometimes days. Demand Spikes The most visible cause of congestion is a sudden surge in demand for block space. These spikes are often triggered by specific events. A popular NFT mint can generate tens of thousands of transactions within minutes as collectors rush to secure limited-edition tokens. A DeFi protocol launch or airdrop claim attracts a flood of users interacting with new smart contracts simultaneously. A major market crash triggers a wave of liquidations, position closures, and panic selling as DeFi participants scramble to adjust their positions. On Ethereum, some of the worst congestion events in the network's history were caused by specific applications. The CryptoKitties craze in December 2017 clogged the network for days. Yuga Labs' Otherside NFT mint in May 2022 caused gas prices to exceed 8,000 gwei, with some users paying thousands of dollars in fees for transactions that ultimately failed. Airdrop claims for tokens like Uniswap's UNI and Arbitrum's ARB created sustained congestion lasting hours as millions of eligible wallets rushed to claim. Meme token trading creates recurring congestion on chains like Solana and Base, where a viral token launch can generate millions of swap transactions within hours. Unlike planned events (where users at least know congestion is coming), meme token surges are unpredictable and can overwhelm network capacity with little warning. Fixed Block Capacity and Production Rate The deeper cause of congestion is that block capacity and production rate are fixed by the protocol's design, while demand is variable and sometimes explosive. Ethereum produces a block every 12 seconds with a gas limit of roughly 30 million gas. A simple ETH transfer costs 21,000 gas. A Uniswap swap costs 150,000-300,000 gas. A complex DeFi interaction might cost 500,000 gas or more. Under ideal conditions, Ethereum can process roughly 15-30 transactions per second depending on the complexity of those transactions. When thousands of users are simultaneously trying to mint an NFT, swap a token, or claim an airdrop, 15-30 TPS is nowhere near enough. Bitcoin faces the same fundamental constraint with even tighter limits. Blocks are produced every 10 minutes (compared to Ethereum's 12 seconds), and each block can hold approximately 2,000-3,000 transactions. This gives Bitcoin a throughput of roughly 5-7 TPS. During periods of high demand, Bitcoin's mempool can accumulate hundreds of thousands of unconfirmed transactions, with low-fee transactions waiting days for confirmation. These capacity limits are not arbitrary. They exist to maintain decentralization. If blocks were larger or produced faster, fewer nodes could keep up with the data processing requirements, which would centralize the network around operators with expensive hardware. The tradeoff between throughput and decentralization is at the heart of blockchain design and is often called the blockchain trilemma (the difficulty of simultaneously achieving scalability, security, and decentralization). MEV and Priority Gas Auctions Maximal Extractable Value (MEV) contributes to congestion in a less obvious but significant way. MEV refers to the profit that block producers and searchers can extract by reordering, inserting, or censoring transactions within a block. When a profitable MEV opportunity appears (such as an arbitrage between two DEXes), multiple searchers compete to capture it by submitting transactions with progressively higher fees. This priority gas auction consumes block space and drives up fees for everyone, even users whose transactions have nothing to do with the MEV opportunity. Sandwich attacks are a specific MEV pattern that directly impacts regular users. A searcher detects a large swap transaction in the mempool, places a buy order before it (front-running) and a sell order after it (back-running), profiting from the price impact the victim's transaction creates. These front-run and back-run transactions consume additional block space and increase congestion beyond what organic user demand would produce. Layer 2 Solutions and Congestion Relief Layer 2 networks (like Arbitrum, Base, Optimism, and zkSync) were designed in part to relieve congestion on Ethereum L1. By processing transactions on a separate chain and posting compressed summaries back to Ethereum, L2s dramatically increase the total throughput available to users. A transaction that costs $5 in gas on Ethereum L1 during congestion might cost $0.01 on an L2. However, L2s can also experience their own congestion during extreme demand spikes, and they inherit some limitations from the L1 they settle to. Solana takes a different approach to congestion management with its local fee markets. Instead of a single global fee market where all transactions compete for the same block space, Solana attempts to isolate congestion to specific "hot" accounts. Transactions interacting with a congested contract (like a popular DEX pool) pay elevated fees, while transactions interacting with unrelated contracts are unaffected. This design reduces the collateral damage of congestion spikes, though it does not eliminate them entirely. How long does blockchain congestion last? The duration of congestion depends on what caused it. A scheduled event like a high-profile NFT mint or an airdrop claim usually produces a sharp spike that clears within minutes to a few hours once the rush of claimants subsides. Sustained congestion is different: when a chain is persistently running near capacity (for example, during a prolonged bull market or a viral meme token cycle), elevated fees and slow confirmations can last days. Bitcoin congestion tends to clear more slowly than Ethereum because its 10 minute block time means the mempool drains in larger, less frequent steps. The practical signal that congestion is easing is a falling mempool size combined with declining priority fees. Congestion is also closely related to why blockchains slow down in general, and it is distinct from a complete outage. If a network stops producing blocks entirely, that is a chain halt, not congestion. How do block limits differ across blockchains? Because every chain caps how much work fits in a block, the headroom before congestion sets in varies widely. The table below summarizes the approximate capacity constraints that determine how quickly each network backs up under load. BlockchainBlock timeCapacity metricApprox. throughputBitcoin~10 minutes4M weight units5 to 7 TPSEthereum L1~12 seconds~30M gas per block15 to 30 TPSSolana~400 ms slotCompute units and account locksThousands of TPSArbitrum / Base (L2)Sub-secondL2 gas, settled to L1Hundreds to thousands of TPS Higher throughput chains and Layer 2 networks absorb demand spikes better, but no design is immune. For a deeper look at the relationship between speed and capacity, see throughput vs latency and what a Layer 2 blockchain is. What is the difference between congestion and a chain halt? Congestion and a chain halt both look like "the network is broken" to an end user, but they are fundamentally different conditions with different causes and fixes. Congestion means the chain is still producing blocks; it is just full and expensive. A chain halt means block production has stopped entirely. AspectCongestionChain haltBlock productionContinues normallyStoppedTransactionsConfirm slowly, high feesDo not confirm at allRoot causeDemand exceeds block capacityConsensus failure or client bugTypical resolutionDemand subsides, fees fallValidators coordinate a restart In short, congestion is a pricing and capacity problem, while a chain halt is an availability problem. MEV activity such as priority gas auctions can intensify congestion, so understanding what MEV is helps explain why fees climb faster than raw user demand alone would predict. How can developers reduce the impact of congestion? Applications cannot eliminate congestion, but they can stay reliable through it. Use dynamic fee estimation rather than hardcoded gas prices, implement retry logic with proper nonce management, and queue non-urgent transactions for off-peak windows. On the data side, prefer push-based delivery over tight polling loops so your ingestion keeps pace as blocks grow. Understanding how RPC requests work, choosing streaming over polling, and accounting for provider RPC rate limiting all reduce the chance that congestion turns into dropped data or failed user transactions. Frequently Asked Questions Why do gas fees spike during congestion? Block space is limited, and the mempool is a fee-based priority queue. When more transactions compete than a block can hold, users bid higher fees to get included first. That bidding war pushes the prevailing gas price up until demand falls back below block capacity. Does congestion mean the blockchain is down? No. During congestion the chain keeps producing blocks and confirming transactions; they are just slower and more expensive. A network that stops producing blocks is experiencing a chain halt, which is a separate problem. Can Layer 2 networks get congested too? Yes. Layer 2s offer far more throughput than Ethereum L1, so they tolerate larger spikes, but extreme demand can still fill their capacity and raise fees. They also inherit some constraints from the Layer 1 they settle to. How can I tell if congestion is clearing? Watch the mempool size and priority fees. A shrinking backlog of unconfirmed transactions combined with falling priority fees is the clearest sign that demand has dropped back below block capacity. Why don't blockchains just make blocks bigger? Larger or faster blocks raise the hardware cost of running a node, which pushes out smaller operators and centralizes the network. Capacity limits are a deliberate tradeoff that protects decentralization, a tension often called the blockchain trilemma. How Congestion Impacts Developers Building on Quicknode For developers, congestion affects both transaction submission and data ingestion. On the submission side, transactions with insufficient gas fees get stuck in the mempool or are dropped entirely. Applications need robust fee estimation, retry logic, and nonce management to handle congested conditions gracefully. Quicknode's Core API provides reliable transaction submission with consistent performance even during peak congestion, and enhanced API methods help applications estimate appropriate gas prices based on current network conditions. On the data ingestion side, congestion means more transactions per block, larger blocks, more events to process, and higher data volumes. Polling-based architectures struggle during congestion because the volume of data per block increases while the polling interval remains fixed, creating a processing backlog. Quicknode Streams handles congestion gracefully because its push-based architecture scales with block size. Whether a block contains 100 transactions or 1,000 transactions, Streams delivers the complete data to your destination with the same guaranteed delivery and ordering. Configurable batching and compression help manage the increased data volume during sustained congestion periods. Further Reading What is a Mempool? - Quicknode Guide Guide to Efficient RPC Requests - Quicknode Getting Started with Streams - Quicknode Docs Quicknode Core API

Question 2

Blockchain throughput vs latency

Accepted Answer

Throughput and latency are two different measurements of blockchain performance that are frequently confused. Throughput is capacity: how many transactions the network can process per second (TPS). Latency is speed: how long it takes for an individual transaction to be confirmed and finalized. A highway analogy helps clarify the distinction. Throughput is how many cars the highway can carry per hour. Latency is how long it takes one car to drive from entrance to exit. A highway can have high throughput (many cars) but high latency (slow speed due to congestion), or low throughput (few cars) but low latency (each car moves fast). The best blockchains optimize for both, but tradeoffs are real.

Question 3

Why blockchains slow down

Accepted Answer

Blockchains slow down when demand for block space exceeds the network's processing capacity. The most common causes are network congestion (too many transactions competing for limited block capacity), state bloat (growing storage requirements that increase the computational cost of processing each block), consensus overhead (communication costs that scale with validator count), and infrastructure limitations (slow nodes, high-latency RPC endpoints, and inefficient data pipelines). Slowdowns manifest as longer confirmation times, higher fees, stale data in applications, and degraded user experience. Some causes are inherent to blockchain design, while others can be mitigated with better infrastructure.

區塊鏈	區段時間	容量指標	約略吞吐量
比特幣	約 10 分鐘	4M 重量單位	每秒 5 至 7 筆交易
以太坊 L1	約 12 秒	每個區塊約 30M 氣體	每秒 15 至 30 筆交易
索拉納	約 400 毫秒的時段	運算單元與帳戶鎖定	數千個 TPS
Arbitrum / Base（L2）	不到一秒	L2 氣體，已結算至 L1	數百至數千個 TPS

面向	交通壅塞	鏈式停止
區塊生產	照常繼續	停止
交易	確認過程緩慢，手續費高昂	完全不要確認
根本原因	需求超過區塊容量	共識失敗或客戶端錯誤
典型解析度	需求減弱，費用下降	驗證者協調系統重啟

想掌握最新動態嗎？

開發者工具

文件與指南

是什麼導致區塊鏈擁塞？

簡單的解釋

需求激增

固定區塊容量與產出率

MEV 與優先天然氣拍賣

第二層解決方案與擁塞緩解

區塊鏈的擁塞狀況會持續多久？

不同區塊鏈的區塊限制有何差異？

「擁塞」與「鏈式中斷」之間有何區別？

開發商該如何減輕交通壅塞的影響？

常見問題

為什麼在網路擁塞時， gas 費用會飆升？

網路擁塞是否代表區塊鏈已停止運作？

第 2 層網路也會發生擁塞嗎？

我該如何判斷鼻塞是否正在緩解？

為什麼區塊鏈不直接把區塊做得更大呢？

擁塞如何影響在 Quicknode 上進行開發的開發者

延伸閱讀

立即開始建置