ecg is a cosmos validators Dead Man's Switch.

---

As long as the monitor is receiving heartbeats from ecg, alerting will not be triggered. This "failure as default" approach makes this kind of monitoring resilient to failures that might otherwise prevent you from receiving any alert.

Examples:

• Your whole infra loses connectivity
• Monitoring services is down at the same time as the validator
• Monitoring service can't access the validator
• Monitoring service lose access to its data source (ie: LCD)

Warning

Monitoring is essential. Ecg and heartbeat style monitoring are complementary to other monitoring tools and methods. We recommend ecg as an aditional failsafe alongside more proactive tools.

There are multiple SaaS platforms like Better Uptime or Dead Man's Snitch that implement heartbeat monitoring.

Features

ecg will:

1. For each target, fetch a state from all clients
2. Keep the freshest one (highest block height)
3. Run checks on it
4. If they all pass, send heartbeat

Checks include:

• block height increased
• validator block misses didn't increase
• validator isn't jailed
• validator isn't tombstoned

Getting Started

This tool is installable as a crate.

cargo install ecg

Or available as a Docker image.

docker pull ghcr.io/setten-io/ecg

Usage

ecg
Heartbeats for cosmos validators

USAGE:
    ecg [OPTIONS]

OPTIONS:
    -h, --help           Print help information
    -p, --path <PATH>    Path to yaml config [env: ECG_CONFIG_PATH=] [default: ecg.yaml]
    -V, --version        Print version information

It is possible to change ecg logging level through the ECG_LOG env var, or for all/any module through the RUST_LOG env var.

Configuration

ecg is configured through a yaml file.

The default path is ./ecg.yaml but can be specified through the -p/--path flag or the ECG_CONFIG_PATH env var.

Example

ecg.yaml

targets:
  phoenix:
    url: https://betteruptime.com/api/v1/heartbeat/fFKHCd3YNkayv8Fr6MJAFE3w
    valcons_address: terravalcons1qqyfhs9oacvteimwdpbt77fis88mie5gx6gxf2
    interval: 10
    clients:
      - type: lcd
        url: https://phoenix-lcd.terra.dev
      - type: lcd
        url: https://terra-api.polkachu.com
      - type: setten-lcd
        project_id: ea08855653b64998bb47b2c03bf66de7
        key: 02215b36969446c28b22059e63b4301b
        network: phoenix
        blockchain: terra
  kaiyo:
    url: https://betteruptime.com/api/v1/heartbeat/t6xm2P7Ujfjz3ph5TNBFti8X
    valcons_address: kujiravalcons14rt55jpahf4giiupxrxivy85ecog2onb29a2ev
    interval: 2
    clients:
      - type: lcd
        url: https://lcd.kaiyo.kujira.setten.io
      - type: lcd
        url: https://kujira-api.polkachu.com

Specification

targets.<name>

The name of your target; Should be self-explanatory.

targets.<name>.url

The url of the monitor to send GET http heartbeat requests to.

targets.<name>.valcons_address

The validator valcons address.

It can be found using the cosmos sdk chain binary cli:

terrad tendermint show-address

targets.<name>.interval

Optional, default to 30

Interval in seconds between each cycle (run checks + send heartbeat).

You must set this in accordance to the heartbeat frequency your monitor is expecting to receive and the chain block time.

targets.<name>.clients

Array of redundant clients to querry in parallel.

targets.<name>.clients[*].type

Type defines the kind of clients and the configuration keys that will be available.

Available client types and their configurations:

lcd

• url - LCD endpoint to query (ex: https://kujira-api.polkachu.com)

setten-lcd

• project_id - Setten project id (ex: ea08855653b64998bb47b2c03bf66de7)
• key - Setten project key (ex: 02215b36969446c28b22059e63b4301b)
• network - Setten network slug (ex: phoenix)
• blockchain - Setten blockchain slug (ex: terra)

For network and blockchain slugs, please see Setten's docs