Skip to main content

Service Provisioning Checlist

Purpose

This checklist is to ensure that all aspects of a new service are provisioned properly, completely, and in the correct order to prevent potential failures elsewhere in the system.

Steps

  • Determine any potential impact to any other services; see things to look out for below
    • Is this service going to be running on app-01 or a different host?
    • Is it going to tie intoutilize SSO auth?
    • Is it going to need a database? Service files folder in /mnt/data/services on app-01?
    • Is it going to need any other secrets?
    • Does this service need to be monitored?
    • Exposed to the public internet?
  • Determine whichthe monitoringmost systemsfeasible needdeployment method
    • Docker container
    • nixOS module (preferred for reproducibility and programmatic configuration)

Check on repology.org to beverify disabled,if permanentlythe andnixOS temporarilymodule is up to preventdate servicewith outageupstream

notifications
  • If this service has a MariaDB, PostgreSQL or otherwise database, remove its entry from the nixOS borgmatic config to prevent backup failure
  • If this service is running in a Docker container, tear down its compose project. Otherwise stop the service and disable/remove its nixOS config. Push configuration change to staging branch

    Do not push this change to main until testing that the configuration builds successfully

  • If this service is a docker-compose project, move its folder to ~/Containers/.retired-services

  • If this service is publicly exposed with a TLS cert, remove its entry from traefik's acme.json file to prevent unwanted cert renewals
  • If remaining data is unwanted, clear all relevant files from the filesystem i.e. /srv/<servicename> and any relevant databases and secrets
  • Push changes fromstaging to main 
  • If any related monitoring systems were temporarily put into maintenance mode, re-enable them in Uptime Kuma and Healthchecks

Vikunja Copy-Paste Version

  • Shutdown/disable needed monitoring services
  • Remove/disable borgmatic database backup entry from nixOS to prevent borgmatic failure
  • Teardown compose project/remove nixOS service config, push change to staging
    • If docker-compose project, move to ~/Containers/.retired-services
  • Remove service's entry from traefik's acme.json file to prevent unwanted cert renewals
  • If unneeded, clear all remaining files from the filesystem i.e. /srv/<servicename> and any relevant databases and secrets
  • Push changes fromstaging to main 
  • Re-enable monitoring systems as needed
Back to top