/restic-robot

Backups done right... by robots! Restic backup but the robot friendly version.

Primary LanguageGoMIT LicenseMIT

Restic Robot

Backups done right... by robots!

This is a small and simple wrapper application for Restic that provides:

  • Automatically creates repository if it doesn't already exist
  • Scheduled backups - no need for system-wide cron
  • Prometheus metrics- know when your backups don't run!
  • JSON logs - for the robots!
  • Pre/post shell script hooks for custom behaviour! (Thanks @opthomas-prime!)

Usage

Just go build and run it, or, if you're into Docker, southclaws/restic-robot.

Environment variables:

  • SCHEDULE: cron schedule
  • RESTIC_REPOSITORY: repository name
  • RESTIC_PASSWORD: repository password
  • RESTIC_ARGS: additional args for backup command
  • RUN_ON_BOOT: run a backup on startup
  • PROMETHEUS_ENDPOINT: metrics endpoint
  • PROMETHEUS_ADDRESS: metrics host:port
  • PRE_COMMAND: A shell command to run before a backup starts
  • POST_COMMAND: A shell command to run if the backup completes successfully
  • TRIGGER_ENDPOINT: manual trigger endpoint

Prometheus metrics:

  • backups_all_total: The total number of backups attempted, including failures.
  • backups_successful_total: The total number of backups that succeeded.
  • backups_failed_total: The total number of backups that failed.
  • backup_duration_milliseconds: The duration of backups in milliseconds.
  • backup_files_new: Amount of new files.
  • backup_files_changed: Amount of files with changes.
  • backup_files_unmodified: Amount of files unmodified since last backup.
  • backup_files_processed: Total number of files scanned by the backup for changes.
  • backup_added_bytes: Total number of bytes added to the repository.
  • backup_processed_bytes: Total number of bytes scanned by the backup for changes

It's that simple!

Manual backups

Sometimes backups are required out-of-band - e.g. before some manual changes to a system are made. Instead of running restic manually, or to edit the cron schedule for a single run, you can trigger a manual backup by sending an HTTP POST request to the configured TRIGGER_ENDPOINT (defaulting to http://localhost:8080/trigger). It reuses the listen address configured with PROMETHEUS_ADDRESS. If the endpoint is set to an empty string, manual backups are disabled.

Docker Compose

Stick this in with your other compose services for instant backups!

services:
  #
  # your stuff etc...
  #

  backup:
    image: southclaws/restic-robot
    restart: always
    environment:
      # every day at 2am
      SCHEDULE: 0 0 2 * * *
      RESTIC_REPOSITORY: my_service_repository
      RESTIC_PASSWORD: ${MY_SERVICE_RESTIC_PASSWORD}
      # restic-robot runs `restic backup ${RESTIC_ARGS}`
      # so this is where you specify the directory and any other args.
      RESTIC_ARGS: /data
      B2_ACCOUNT_ID: ${B2_ACCOUNT_ID}
      B2_ACCOUNT_KEY: ${B2_ACCOUNT_KEY}
    volumes:
      # Bind whatever directories to the backup container.
      # You can safely bind the same directory to multiple containers.
      - "/container_data/blog/wordpress:/data/wordpress"