Blockscout explorer for Somnia's high-throughput L1
Protofire deployed and tuned Somnia's Blockscout explorer so it indexes the chain in near real time. At peak it handled 103 million transactions in a single day, backed by a PostgreSQL instance tuned to hold more than 2 billion rows and 19-plus terabytes without query slowdown. The live explorer is at shannon-explorer.somnia.network.
“If the explorer cannot ingest and serve data as fast as the chain produces it, the network looks opaque no matter how fast the base layer runs.”
Why a fast Layer 1 still needs an explorer that can keep up
Somnia is an EVM-compatible Layer 1 that the Somnia team built for very high throughput: bursts of up to one million transactions per second and sub-second finality (source: Somnia). During its testnet phase the network sustained heavy daily volume, and every transaction is a block, contract call, or transfer that someone eventually needs to look up.
A block explorer is the public window into all of it: contract verification, transaction and block search, address history, and the API that ecosystem tools read from. If the explorer cannot ingest and serve data as fast as the chain produces it, the network looks opaque no matter how fast the base layer runs.
Why the throughput broke a default explorer setup
Blockscout is a capable open-source explorer, but its default configuration is tuned for ordinary L1 and L2 volumes, not for a chain pushing toward a million transactions per second. Three things broke under Somnia's load. First, ingestion: indexing had to keep up with block production without falling permanently behind.
Second, the database: a PostgreSQL instance growing past 2 billion rows and 19 terabytes started to show query slowdowns and indexing lag, exactly when fast lookups mattered most. Third, the machine and batch settings underneath were sized for normal traffic, so contract verification, search, and analytics degraded together under peak load.
How Protofire deployed and optimized Somnia's Blockscout explorer
Protofire built Somnia's explorer on Blockscout's open-source foundation and then did the work that a stock deployment skips. On the database side, we tuned PostgreSQL for large-scale ingestion, adjusting memory handling and vacuum thresholds and adding the indexes and query rewrites that fit Somnia's workload, so a dataset past 2 billion rows kept serving lookups without slowing down.
On the infrastructure side, we deployed a 96-core instance with 1 TB of RAM and more than 100 TB of storage under containerized orchestration, and tuned batch sizes and concurrency so the indexer stayed close to the head of a high-throughput chain instead of drifting behind. Throughout, we kept the core explorer features working at that scale: contract verification, lookups, block search, and the API that partner protocols and ecosystem tools depend on.
Results
The figures below come from the Somnia explorer Protofire deployed and operated during the engagement.
The verifiable artifact is the live explorer at shannon-explorer.somnia.network, which serves Somnia's on-chain data in near real time. There is no public repo for the explorer work: Blockscout is upstream and the tuning is operational.
Protofire's broader engineering engagement with the Somnia ecosystem is public in the Protofire GitHub org: the audit reports in Somnialend-Audits and the Chainlink-compatible price-feed work in somnia-price-feeds are separate, verifiable evidence of the relationship, not part of the explorer.