v25.01

NEWS:
  - run custom command at each host
  - prune git repo if desired

CHANGES:
  - add ServerAliveInterval=60s
  - move more Tor config options to the inventory

v24.12

NEWS:
  - grep proxy log for NAT type
  - add hourly cronjob to ensure that service is up
  - force daemon restart by tag "force-restart"

CHANGES:
  - ASH and HIL are too expensive to be selected by default
  - remove throttling of writing to local files

v24.11

CHANGES:
  - enhance pattern for issues

v24.10.2

CHANGES:
  - restrict mem access for self-compiled kernels

v24.10.1

FIXES:
  - put missing file under source code control

v24.10

NEWS:
  - get bridge stats
  - add reaper cron job for hanging apt update
  - reboot if kernel was upgraded and not yet rebooted

CHANGES:
  - use mincoredumper instead default one
  - do not throttle github.com
  - renamed "local" inventory file to allow overwriting its values easily

v24.09

NEWS:
  - poweron role

CHANGES:
  - factored out base setup and CA
  - for self-compiled kernel: do not start app before the very first kernel booted
  - use only HTTPs Debian repositories
  - Snowflake metrics intervall shorten from 1 min to 15 sec
  - no_log is default

v24.08

NEWS:
  - reduce TCP FIN timeout to 10 sec
  - set hostname
  - add daily cron job for server to update the Tor authority ip adresses
  - install Debian cloud kernel
  - async build of kernel
  - info task: issue
  - start to re-org playbooks
  -

CHANGES:
  - deactivate all stats for Tor relays till
    https://gitlab.torproject.org/tpo/core/tor/-/issues/40958 is solved
  - configure SSH in a more safe manner

v24.07

NEWS:
  - base firewall: allow incoming DHCP
  - use torutils metrics script for tor server

CHANGES:
  - no swap file
  - Tor info log is rotated hourly
  - print etime instead lstart in uptime
  - use KTLS in NGinx
  - derive MyFamily from wellknown file

v24.06

NEWS:
  - use external firewall script for Tor server
  - handle Hetzner cx11 type too (1 vCPU Intel)

CHANGES:
  - use chrony instead of ntp for time sync
  - handle more than 1 prometheus server