Skip to content

Strong degrading performance during continuous use of Electrs #1169

@MvdS74

Description

@MvdS74

I'm running Electrs + Bitcoin Core on a dedicated Debian machine (Xeon E3-1225V6 4c/4t, 64GB DDR4, 2TB m.2 NVME SSD).

The startup information given by Electrs is: Mar 06 10:35:11 Hephaistos electrs[567]: Starting electrs 0.10.9 on x86_64 linux with Config { network: Bitcoin, db_path: "/var/lib/electrs/db/bitcoin", db_log_dir: None, db_parallelism: 4, daemon_auth: CookieFile("/var/lib/bitcoin/.bitcoin/.cookie"), daemon_rpc_addr: 127.0.0.1:8332, daemon_p2p_addr: 127.0.0.1:8333, electrum_rpc_addr: 127.0.0.1:50001, monitoring_addr: 127.0.0.1:4224, wait_duration: 10s, jsonrpc_timeout: 15s, index_batch_size: 10, index_lookup_limit: None, reindex_last_blocks: 0, auto_reindex: true, ignore_mempool: false, sync_once: false, skip_block_download_wait: false, disable_electrum_rpc: false, server_banner: "Welcome to electrs 0.10.9 (Electrum Rust Server)!", signet_magic: f9beb4d9 }

After starting up, I see that Electrs is capable of handling around 250 calls per seconds (coming from multiple clients/threads) and that seems to be degrading almost immediately until it stabilizes at around 70-80 calls per second. CPU utilization is stable around 30-35%, ram utilization around 10%, SSD write around 0-5 IOPS, read around 50-100 IOPS.

This information shows me that the system is not close to being maxed out, so I think that Electrs is struggling.

Can I test this somehow? Can I do optimalisations in the config to get Electrs to perform better?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions