Skip to main content

Network Time to Consensus

This metric measures the SecC2RC (Seconds from Consensus to Record Creation) time, which represents the elapsed time from when a transaction reaches consensus until its corresponding record is created and made available. This statistic serves as a proxy for network post-consensus processing efficiency and overall system latency on the Hedera network.

Hedera Data Access

To access this Hedera network statistic (and others) via Hgraph's GraphQL & REST APIs, create a free account at app.hgraph.com.

Incremental update setup

This Hedera Stat requires Prometheus and access to Hedera telemetry data. Please refer to the installation guide for more information.

Hedera Stat Name: avg_time_to_consensus

Methodology​

Data Source​

The avg_time_to_consensus metric is derived from the Hedera network's Prometheus telemetry system, specifically using the platform_secC2RC{environment="mainnet"} metric. This provides real-time measurements of the SecC2RC (Seconds from Consensus to Record Creation) timing.

ETL Pipeline Architecture​

The metric collection follows a three-stage ETL (Extract, Transform, Load) pipeline:

  1. Extract: Queries Prometheus using promtool to retrieve platform_secC2RC data
  2. Transform: Processes JSON response to CSV format with proper timestamp handling
  3. Load: Bulk inserts data into PostgreSQL ecosystem.metric table

Data Processing​

Time Period Support​

  • Hourly: Collects data every hour using --step=1h in Prometheus queries
  • Daily: Collects daily aggregates using --step=1d in Prometheus queries

Timestamp Precision​

  • Maintains Hedera's nanosecond precision by converting Unix timestamps to nanoseconds
  • Uses PostgreSQL's int8range type for timestamp intervals
  • Handles proper time boundary alignment (hour/day boundaries)

Calculation Formula​

SecC2RC Time = Record Creation Timestamp - Consensus Timestamp

The raw Prometheus metric provides this calculation; the ETL pipeline aggregates these values over the specified time periods (hourly or daily averages).

Automated Collection​

Data collection is automated via system cron jobs:

  • Hourly: Runs every hour to collect recent data
  • Daily: Runs daily at 00:02 UTC for daily aggregates
  • Backfill: Supports historical data collection in yearly chunks (2023-2025)

GraphQL API Examples​

Test out these queries using our developer playground.

Fetch most recent network time to consensus (hour)​

query GetRecentNetworkTTC {
  ecosystem_metric(
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "hour" } }
    order_by: { end_date: desc_nulls_last }
    limit: 1
  ) {
    total
    end_date
  }
}

Fetch hourly average TTC (timeseries)​

query HourlyNetworkTTC {
  ecosystem_metric(
    order_by: { end_date: desc_nulls_last }
    limit: 8760
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "hour" } }
  ) {
    total
    end_date
  }
}

Fetch most recent network time to consensus (day)​

query GetRecentNetworkTTCDaily {
  ecosystem_metric(
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "day" } }
    order_by: { end_date: desc_nulls_last }
    limit: 1
  ) {
    total
    end_date
  }
}

Fetch daily average TTC (timeseries)​

query DailyNetworkTTC {
  ecosystem_metric(
    order_by: { end_date: desc_nulls_last }
    limit: 365
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "day" } }
  ) {
    total
    end_date
  }
}

7 day percentage change (period comparison)​

query TTC7DayChange {
  current: ecosystem_metric_aggregate(
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "hour" } }
    order_by: { start_date: desc_nulls_last }
    limit: 168
  ) {
    aggregate {
      avg {
        total
      }
    }
  }
  previous: ecosystem_metric_aggregate(
    where: { name: { _eq: "avg_time_to_consensus" }, period: { _eq: "hour" } }
    order_by: { start_date: desc_nulls_last }
    offset: 168
    limit: 168
  ) {
    aggregate {
      avg {
        total
      }
    }
  }
}

Available Time Periods​

The period field supports the following values:

  • hour
  • day

ETL Implementation​

The Time to Consensus metric is implemented using an ETL pipeline rather than SQL functions. The implementation files are located in the Hedera Stats GitHub repository:

ETL Pipeline Components:

  • Extract: src/time-to-consensus/etl/extract.sh (hourly) and src/time-to-consensus/etl/extract_day.sh (daily)
  • Transform: src/time-to-consensus/etl/transform.sh (hourly) and src/time-to-consensus/etl/transform_day.sh (daily)
  • Load: src/time-to-consensus/etl/load.sh and src/time-to-consensus/etl/load.sql
  • Orchestration: src/time-to-consensus/run.sh (hourly) and src/time-to-consensus/run_day.sh (daily)
  • Documentation: src/time-to-consensus/CRON_SETUP.md

View GitHub Repository →

Dependencies​