updog

README

updog

Updog is a health check aggregator for scenarios where you have multiple micro services running in a group. For example you may have many data centres running the same services. Your geo-steered loadbalancer could hit updog's /health or /updog endpoint for the status of each data centre. Updog currently supports health checks against the following dependencies:

• HTTP
• Redis
• MSSQL
• PostgreSQL

Checks of dependencies are executed in parallel, to ensure one slow dependency doesn't risk causing an upstream timeout. The status of each dependency check is returned for adhoc debugging. Any response between 200 ≤ x ≤ 299 will succeed. /health and /updog return a 502 if any dependencies fail. Sample response:

{
 "GitHub": {
  "type": "http",
  "success": true,
  "duration": 0.4550456,
  "Reason": null
 },
 "Google": {
  "type": "http",
  "success": true,
  "duration": 0.6515392,
  "Reason": null
 },
 "Prometheus": {
  "type": "http",
  "success": true,
  "duration": 0.6485875,
  "Reason": null
 }
}

Configuration

Configuration of the downstream dependencies is done with a simple YAML file.

dependencies:
  http:
  - name: Google
    http_endpoint: "https://google.com"
  - name: GitHub
    http_endpoint: "https://github.com"
  redis:
  - name: Redis with SSL
    address: "foo.redis.cache.windows.net:6380"
    password: securePassword
    ssl: true
  - name: Insecure Instance
    address: "localhost:6379"
    password: ""
    ssl: false  
  sql:
  - name: Microsoft SQL Database
    type: mssql
    # Template sqlserver://username:password@db-server:port?paramName=paramValue
    connectionString: "sqlserver://username:password@foo.database.windows.net:1433?database=dbname"
  - name: Posrgres Database
    type: postgres
    # Template postgres://user@host:password@DBserver/databaseName?paramName=paramValue
    connectionString: "postgres://user@foo-postgres:password@foo-postgres.postgres.database.azure.com/dbname?sslmode=verify-full"

Flags

The following flags can be supplied on startup.

Flags:
      --help                    Show context-sensitive help (also try --help-long and --help-man).
  -t, --timeout=5s              Timeout for dependency checks
  -c, --config.path="updog.yaml"
                                Path of configuration file
      --listen.address=":1111"  Address to listen on for HTTP requests
      --version                 Show application version.

Docker

Docker images available at Docker Hub

Metrics

Metrics are exposed in the Prometheus format, at the standard /metrics endpoint.

Name	Description	Type	Labels
updog_http_request_duration_seconds	Inbound request latency.	Histogram	path
updog_dependency_duration_seconds	Latency of the outbound dependency check.	Histogram	dependency, dependency_type
updog_dependency_checks_total	Count of total health checks per dependency.	Counter	dependency, dependency_type
updog_dependency_check_failures_total	Count of total health check failures per dependency.	Counter	dependency, dependency_type