Commit graph

123 commits

Author SHA1 Message Date
5f4b9c6bed Update docker.io/library/redis Docker tag to v8
Some checks failed
/ Ansible Lint (push) Failing after 47s
/ Ansible Lint (pull_request) Failing after 47s
2026-01-25 20:45:50 +00:00
c7d51af5b4
rollout Alloy to replace prometheus_node_exporter
With the new network we need to deploy a push based solution in order to get metrics into prometheus
2026-01-25 21:44:49 +01:00
11779ab21d
grafana: get alertmanager to be more chill
Some checks are pending
/ Ansible Lint (push) Waiting to run
a bit of help to deal with alert fatigue
2026-01-25 21:41:20 +01:00
0939771d08
public-reverse-proxy(host): add entries for cpu.ccc.de
Some checks failed
/ Ansible Lint (push) Failing after 43s
2026-01-25 20:22:44 +01:00
cee1fe970a Add spaceapiccc as a replacement for erfafoo
Some checks failed
/ Ansible Lint (push) Failing after 42s
2026-01-25 14:03:54 +01:00
d514688574
systemd_networkd(role),router(host): support global config to fix forw.
All checks were successful
/ Ansible Lint (push) Successful in 1m58s
With the router upgrade to Debian 13 the systemd version got upgraded as
well breaking the current configuration for IP forwarding.
Add a variable for global systemd-networkd configuration and use that to
enable IPv4 and IPv6 forwarding on the router.

The systemd_networkd role could be a bit nicer, not deploying/deleting
the global configuration, if the variable is empty and
reloading/restarting systemd-networkd at appropriate times. But as is
works for now.
2026-01-18 19:21:33 +01:00
428b5c70bc
pretalx(host): roll back to pretalx v2025.1.0 for celery as well 2026-01-13 14:19:57 +01:00
3e0fdfa8de
pretalx(host): roll back to pretalx v2025.1.0 as v2025.2.2 doesn't work
Some checks failed
/ Ansible Lint (push) Failing after 1m56s
2026-01-13 03:43:28 +01:00
c638790819 Update all stable non-major dependencies
Some checks failed
/ Ansible Lint (pull_request) Failing after 2m31s
/ Ansible Lint (push) Failing after 2m5s
2026-01-12 02:30:47 +00:00
968e29ccb8
do v6-only for internal proxy protocol communication
Some checks failed
/ Ansible Lint (push) Failing after 2m5s
Since we want to do v6-only internally, only listen on v6 for proxy
protocol.
This is also needed as we only have set_real_ip_from pointing to a v6.
2026-01-12 03:02:09 +01:00
255327952e
ntfy(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m59s
2026-01-11 03:57:11 +01:00
1971598e71
pretalx(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m55s
2026-01-11 03:23:18 +01:00
372f264bcb
ccchoir(host): move to new network and hostname 2026-01-11 03:23:14 +01:00
2fbb37db18
grafana(host): move to new network and hostname 2026-01-11 03:23:01 +01:00
bb30e88404
router(host): allowlist only certain icmpv6 types
Some checks failed
/ Ansible Lint (push) Failing after 2m14s
2026-01-11 00:29:16 +01:00
a41b07949c
zammad(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m56s
2026-01-11 00:22:37 +01:00
ff550cbd8a
tickets(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 2m22s
2026-01-11 00:00:18 +01:00
49e3ecb986
netbox(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 2m3s
2026-01-09 03:05:29 +01:00
40b67c6bc3
sunders(host): move to new network and hostname 2026-01-07 18:46:16 +01:00
80ddb2efc9
router: enable a DHCP server for the v4-NAT network as well
As the hosts don't really need a static v4, just do DHCP.
2026-01-07 17:25:27 +01:00
944c8cde82
onlyoffice(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 2m5s
2025-12-17 03:34:39 +01:00
366456eff8
keycloak(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m56s
Also just listen on port 8443 for keycloak-admin proxy protocol.
2025-12-16 21:50:40 +01:00
1ca71a053e
pad(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m57s
2025-12-16 21:12:21 +01:00
570600fce3
eh22-wiki(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 1m59s
2025-12-16 20:58:05 +01:00
5a476f2103
cloud(host): move to new network and hostname
Some checks failed
/ Ansible Lint (push) Failing after 2m0s
2025-12-16 20:47:44 +01:00
b72dee0d6d
wiki(host): actually have nginx listen on v6
Some checks failed
/ Ansible Lint (push) Failing after 1m58s
2025-12-16 19:52:24 +01:00
8b94a49f5e
wiki(host): move to new network and internal hostname
Some checks failed
/ Ansible Lint (push) Failing after 2m2s
2025-12-16 19:23:33 +01:00
5f98dca56c
router(host): expose public v6 networks
Also prepare for exposing public v4 networks later.
2025-12-16 19:03:36 +01:00
66ee44366b public-reverse-proxy: New IP of wiki VM 2025-12-14 15:39:03 +01:00
183b91b9f2
router(host): add nftables config for basic router functionality
Some checks failed
/ Ansible Lint (push) Failing after 1m56s
2025-12-13 22:07:38 +01:00
a9e394da06
router(host): add systemd-networkd-based network config 2025-12-13 22:07:37 +01:00
0eaaf9227c Update all stable non-major dependencies
Some checks failed
/ Ansible Lint (pull_request) Failing after 2m4s
/ Ansible Lint (push) Failing after 2m6s
2025-11-19 13:30:39 +00:00
ddab157600
don't pin digests anymore
Some checks failed
/ Ansible Lint (push) Failing after 47s
The benefit of digest pinning isn't that great for this project really
and it comes at the cost of more issues and additional renovate noise,
so just don't anymore.
Adjust renovate config accordingly as well.
2025-11-18 14:24:21 +01:00
80acd5fdc6
grafana: store date for up to 28 days
Some checks failed
/ Ansible Lint (push) Failing after 1m58s
2025-11-11 23:03:59 +01:00
6fea98ffd2 Redirect to 39c3 instead of eh22
Some checks failed
/ Ansible Lint (push) Failing after 2m3s
2025-11-07 20:09:02 +01:00
c3f71b1f08 sunders: replace password in healthcheck with dynamic secret
Some checks failed
/ Ansible Lint (push) Failing after 2m2s
#55

Co-authored-by: ViMaSter <vincent@mahn.ke>
Co-committed-by: ViMaSter <vincent@mahn.ke>
2025-11-02 20:24:55 +01:00
dc6c7cbfb7
sunders(host): deploy sunders using docker compose
Some checks failed
/ Ansible Lint (push) Failing after 2m29s
https://git.hamburg.ccc.de/CCCHH/sunders
2025-11-01 17:53:08 +01:00
a11ccaf16c
disable digest pinning for our images, since Forgejo cleans them up
Some checks failed
/ Ansible Lint (push) Failing after 1m59s
Since Forgejo seems to clean up older tag versions, so older digests,
disable digest pinning for our images.
While generally resulting in undeployable config, with ansible-pull the
breakage is especially noticeable.
2025-10-30 05:50:42 +01:00
a03ed9a362 Update all stable non-major dependencies
Some checks failed
/ Ansible Lint (pull_request) Failing after 2m33s
/ Ansible Lint (push) Failing after 2m6s
2025-10-29 23:45:38 +00:00
01890fecbd Update all digest updates
Some checks failed
/ Ansible Lint (pull_request) Failing after 2m29s
/ Ansible Lint (push) Failing after 2m22s
2025-10-29 00:30:36 +00:00
df32e1cac8
zammad(host): move to latest upstream compose file version (v14.1.1)
Some checks failed
/ Ansible Lint (pull_request) Failing after 2m11s
/ Ansible Lint (push) Failing after 2m6s
It hopefully fixes bugs we had in the past, so removing the workarounds
and it also comes with default values now, so removing all variables set
to those defaults.
2025-10-24 22:57:03 +02:00
747e5b2d4c
zammad(host): change to .env configuration
Some checks failed
/ Ansible Lint (push) Failing after 2m8s
/ Ansible Lint (pull_request) Failing after 2m7s
Align the compose.yaml to upstreams v11.2.0 version.
This is a first step to hopefully then just use the upstreams version
directly and not vendor it.
2025-10-24 22:15:48 +02:00
2f8897751b Pin dependencies
All checks were successful
/ Ansible Lint (pull_request) Successful in 1m59s
/ Ansible Lint (push) Successful in 2m4s
2025-10-23 12:15:38 +00:00
43fac32424
use full image source and set version for all Chaosknoten compose images
Some checks failed
/ Ansible Lint (push) Failing after 52s
Use a full image source and set and explicit version for all images in
Chaosknoten (docker-)compose files.
With Renovate now set up, it is feasible to explicitly set versions.
2025-10-21 23:59:56 +02:00
f3902b43b1
update keycloak and nextcloud
Some checks failed
/ Ansible Lint (push) Failing after 59s
2025-10-20 21:11:28 +02:00
425d302fa9
zammad(host): fix compose by using different elasticsearch image
Some checks failed
/ Ansible Lint (push) Failing after 50s
Also execute manual update steps as outlined here:
https://github.com/zammad/zammad-docker-compose/releases/tag/v14.0.0

Note that this is updating only a part of the compose file and other
parts still are out-of-date and should be updated (like the Postgres).
2025-10-20 01:36:00 +02:00
8a8fdf5f97
renovate(host): set platform to forgejo 2025-10-17 01:26:18 +02:00
4b1c64b615
renovate(host): remove deprecated and no longer working logLevel option
Some checks failed
/ Ansible Lint (push) Failing after 52s
Remove deprecated and no longer working logLevel config option. Debug
output can be enabled by using the LOG_LEVEL environment variable on
demand.
2025-10-16 18:08:29 +02:00
e76c66d74c
renovate(host): set up Renovate using role and exisiting config
Some checks failed
/ Ansible Lint (push) Failing after 56s
Set up Renovate using role and existing config from current
Forgejo/GitHub Actions-based deployment.
2025-10-16 17:52:30 +02:00
1322bcec58
reverse proxy configuration for element-admin
Some checks failed
/ Ansible Lint (push) Failing after 47s
2025-10-12 20:20:02 +02:00