Add local rspamd and redis to the mail edge gateway (#553)

Edge gateways need cheap deterministic in-session filtering behind the postfix milter, with no AI and no ClamAV on 1GB VMs.

- add a signature-checked rspamd yum repo via artifactapi, enabled for the gateway role
- add profiles::rspamd::edge managing the package and local.d files from hiera
- fetch rule files without blocking the service: empty placeholder, swap only on a successful changed fetch, then reload
- run a self-scanning milter proxy on 11332; reject at 13, header at 6, no greylisting, subject rewrite or gpt
- add a local 128MB redis for bayes autolearn; load SpamAssassin rules from rspamd-rules
- bypass rspamd for mail relayed out by the smarthost

Stacked on #552; requires terraform-artifactapi #63 applied first.

Reviewed-on: #553
Co-authored-by: unkin-agent <unkin-agent@unkin.net>
Co-committed-by: unkin-agent <unkin-agent@unkin.net>

hieradata/os/AlmaLinux/all_releases.yaml

@@ -147,6 +147,18 @@ profiles::yum::global::repos:
     baseurl: https://artifactapi.k8s.syd1.au.unkin.net/api/v1/remote/postgresql/common/redhat/rhel-%{facts.os.release.major}-%{facts.os.architecture}
     gpgkey: https://artifactapi.k8s.syd1.au.unkin.net/api/v1/remote/postgresql/keys/PGDG-RPM-GPG-KEY-RHEL
 
+  # Rspamd repositories
+  rspamd:
+    ensure: absent
+    name: rspamd
+    descr: rspamd repository
+    target: /etc/yum.repos.d/rspamd.repo
+    baseurl: https://artifactapi.k8s.syd1.au.unkin.net/api/v1/remote/rspamd/rpm-stable/centos-%{facts.os.release.major}/x86_64
+    gpgkey: https://artifactapi.k8s.syd1.au.unkin.net/api/v1/remote/rspamd/rpm/rspamd.asc
+    gpgcheck: 1
+    repo_gpgcheck: 1
+    mirrorlist: absent
+
   # Ceph repositories
   ceph:
     ensure: absent

hieradata/roles/infra/mail/gateway.yaml

@@ -1,4 +1,58 @@
 ---
+hiera_include:
+  - redis
+
+profiles::yum::global::repos:
+  rspamd:
+    ensure: present
+
+# local redis for rspamd bayes and history; bayes tokens carry a ttl so volatile-ttl can evict
+redis::maxmemory: '128mb'
+redis::maxmemory_policy: 'volatile-ttl'
+
+# local rspamd, deterministic checks only; module defaults cover spf/dkim/dmarc/arc/rbl/phishing/mime_types/fuzzy_check
+profiles::rspamd::edge::rules:
+  edge.cf: 'https://git.unkin.net/unkin/rspamd-rules/raw/branch/main/edge.cf'
+profiles::rspamd::edge::local_d:
+  worker-proxy.inc: |
+    bind_socket = "127.0.0.1:11332";
+    milter = yes;
+    timeout = 120s;
+    count = 1;
+    upstream "local" {
+      default = yes;
+      self_scan = yes;
+    }
+  worker-normal.inc: |
+    enabled = false;
+  redis.conf: |
+    servers = "127.0.0.1:6379";
+  actions.conf: |
+    reject = 13;
+    add_header = 6;
+    rewrite_subject = null;
+    greylist = null;
+  greylist.conf: |
+    enabled = false;
+  gpt.conf: |
+    enabled = false;
+  classifier-bayes.conf: |
+    backend = "redis";
+    expire = 8640000;
+    autolearn {
+      spam_threshold = 6.0;
+      ham_threshold = -0.5;
+      check_balance = true;
+    }
+  spamassassin.conf: |
+    ruleset = "/etc/rspamd/rules/*.cf";
+    alpha = 0.1;
+  settings.conf: |
+    smarthost_outbound {
+      priority = high;
+      ip = "103.216.191.185";
+      want_spam = yes;
+    }
 
 # base postfix configuration (passed to postfix class)
 postfix::relayhost: 'direct'

site/profiles/manifests/rspamd/edge.pp

@@ -0,0 +1,68 @@
+# local rspamd milter for the mail edge gateways
+# - local_d: /etc/rspamd/local.d file name => content
+# - rules: /etc/rspamd/rules file name => url, re-fetched every puppet run; fetch failures keep the last good copy
+class profiles::rspamd::edge (
+  Hash[String, String]           $local_d = {},
+  Hash[String, Stdlib::HTTPSUrl] $rules   = {},
+) {
+
+  package { 'rspamd':
+    ensure => installed,
+  }
+
+  $local_d.each |$name, $content| {
+    file { "/etc/rspamd/local.d/${name}":
+      ensure  => file,
+      owner   => 'root',
+      group   => 'root',
+      mode    => '0644',
+      content => $content,
+      require => Package['rspamd'],
+      notify  => Service['rspamd'],
+    }
+  }
+
+  file { '/etc/rspamd/rules':
+    ensure  => directory,
+    owner   => 'root',
+    group   => 'root',
+    mode    => '0755',
+    require => Package['rspamd'],
+  }
+
+  $rules.each |$name, $url| {
+    $path = "/etc/rspamd/rules/${name}"
+
+    # empty placeholder so rspamd starts before the first successful fetch
+    file { $path:
+      ensure  => file,
+      owner   => 'root',
+      group   => 'root',
+      mode    => '0644',
+      content => '',
+      replace => false,
+    }
+
+    # a failed or unchanged fetch skips the swap without failing the run
+    exec { "rspamd_rules_fetch_${name}":
+      command => "/usr/bin/mv -f ${path}.new ${path}",
+      onlyif  => "/usr/bin/curl -fsS --max-time 30 -o ${path}.new ${url} && ! /usr/bin/cmp -s ${path}.new ${path}",
+      require => File[$path],
+      notify  => Exec['rspamd_reload'],
+    }
+  }
+
+  # reload keeps the milter answering while config changes apply
+  service { 'rspamd':
+    ensure  => running,
+    enable  => true,
+    restart => '/usr/bin/systemctl reload rspamd',
+    require => Package['rspamd'],
+  }
+
+  exec { 'rspamd_reload':
+    command     => '/usr/bin/systemctl reload rspamd',
+    refreshonly => true,
+    require     => Service['rspamd'],
+  }
+}

site/roles/manifests/infra/mail/gateway.pp

@@ -6,6 +6,7 @@ class roles::infra::mail::gateway {
   }else{
     include profiles::defaults
     include profiles::base
+    include profiles::rspamd::edge
     include profiles::postfix::gateway
   }
 }