diff --git a/roles/ai3-prometheus/templates/rules/alerts_auth.conf.yml b/roles/ai3-prometheus/templates/rules/alerts_auth.conf.yml
index 30b99ce8820ded7a629fcabbd39ae61bbf54decb..4f05a9017adc50f99a85fd178793981533294d1b 100644
--- a/roles/ai3-prometheus/templates/rules/alerts_auth.conf.yml
+++ b/roles/ai3-prometheus/templates/rules/alerts_auth.conf.yml
@@ -17,7 +17,7 @@ groups:
annotations:
summary: 'Too many login failures for service {{$labels.exported_service}}'
description: 'The percentage of successful authentications on auth-server with service={{$labels.exported_service}} is too low ({{$value}}). This can indicate a brute-forcing attack (depending on the service), or a failure in the auth-server itself.'
- runbook: '[[ alert_playbook_url ]]/TooManyLoginFailures.md'
+ runbook: '[[ alert_playbook_url ]]/TooManyLoginFailures'
# We also want to check that, for some important services, there
# actually are any successful logins. The threshold is low on
@@ -35,5 +35,5 @@ groups:
annotations:
summary: 'No successful logins for service {{$labels.exported_service}}'
description: 'The auth-server is not reporting successful logins with service={{$labels.exported_service}}. This might indicate something broken with the auth-server itself.'
- runbook: '[[ alert_playbook_url ]]/NoLogins.md'
+ runbook: '[[ alert_playbook_url ]]/NoLogins'
diff --git a/roles/ai3-prometheus/templates/rules/alerts_postfix.conf.yml b/roles/ai3-prometheus/templates/rules/alerts_postfix.conf.yml
index d5c98c7ebcfbf13e7cf1ad0fd4d959e8400702e0..eb07a1e9216bc574d90d9ca7340d564966086165 100644
--- a/roles/ai3-prometheus/templates/rules/alerts_postfix.conf.yml
+++ b/roles/ai3-prometheus/templates/rules/alerts_postfix.conf.yml
@@ -29,7 +29,7 @@ groups:
annotations:
summary: "Postfix is deferring many messages on {{$labels.postfix_instance}}"
description: "The Postfix instance {{$labels.postfix_instance}} is unexpectedly deferring lots of messages. Perhaps some of the expected destinations are unreachable."
- runbook: "[[ alert_playbook_url ]]/PostfixHighDeferred.md"
+ runbook: "[[ alert_playbook_url ]]/PostfixHighDeferred"
- alert: PostfixHighBounces
expr: 'instance:smtp_bounces:rate10m{postfix_instance!="postfix-out"} > 2'
@@ -40,7 +40,7 @@ groups:
annotations:
summary: "Postfix is bouncing many messages on {{$labels.postfix_instance}}"
description: "The Postfix instance {{$labels.postfix_instance}} is unexpectedly bouncing lots of messages. Perhaps there are issues with the user databases."
- runbook: "[[ alert_playbook_url ]]/PostfixHighBounces.md"
+ runbook: "[[ alert_playbook_url ]]/PostfixHighBounces"
# Warn if there are many rejects on an instance: something may (or may not) be broken.
- alert: PostfixHighRejects
@@ -63,7 +63,7 @@ groups:
annotations:
summary: "Postfix is rejecting many messages on {{$labels.postfix_instance}}"
description: "The Postfix instance {{$labels.postfix_instance}} is rejecting a high number of messages. Could be a spam attack, or something broken in the address delivery (if the failing deliveries are for real emails), which is often a DNS problem."
- runbook: "[[ alert_playbook_url ]]/PostfixHighRejects.md"
+ runbook: "[[ alert_playbook_url ]]/PostfixHighRejects"
# Page if there are multiple Postfix instances with high rejects: stuff is really broken.
- alert: PostfixHighRejects
@@ -75,7 +75,7 @@ groups:
annotations:
summary: "Postfix is rejecting many messages"
description: "Multiple Postfix instances are rejecting a high number of messages. There is most likely something wrong with our internal email delivery."
- runbook: "[[ alert_playbook_url ]]/PostfixHighRejects.md"
+ runbook: "[[ alert_playbook_url ]]/PostfixHighRejects"
# Alert when many emails are ratelimited (likely an account compromise)
- alert: PostfixHighRatelimit
diff --git a/roles/ai3-prometheus/templates/rules/alerts_security.conf.yml b/roles/ai3-prometheus/templates/rules/alerts_security.conf.yml
new file mode 100644
index 0000000000000000000000000000000000000000..d0126f3b88679fd49e41cdbe97ea6297c347632a
--- /dev/null
+++ b/roles/ai3-prometheus/templates/rules/alerts_security.conf.yml
@@ -0,0 +1,18 @@
+groups:
+ - name: roles/ai3-prometheus/files/rules/alerts_security.conf
+ rules:
+
+ - alert: UnexpectedConnections
+ expr: sum(delta(unexpected_connections{job="hark"}[1m]) by (host)) > 0
+ for: 1m
+ labels:
+ severity: page
+ scope: host
+ annotations:
+ runbook: '[[ alert_playbook_url ]]/UnexpectedConnections'
+ summary: '[SECURITY] Unexpected connections to {{$labels.host}}'
+ description: |
+ The 'hark' canary listener has detected unexpected connections
+ on host {{$labels.host}}, this could be a sign of an intruder
+ running a port scan on the internal network.
+
diff --git a/roles/ai3-prometheus/templates/rules/alerts_service_prober.conf.yml b/roles/ai3-prometheus/templates/rules/alerts_service_prober.conf.yml
index 39a55d65e768b723ca00221cffb090b9cb04f0fc..8830c3c3c94c67e0f54bd3f063d011db80de5d63 100644
--- a/roles/ai3-prometheus/templates/rules/alerts_service_prober.conf.yml
+++ b/roles/ai3-prometheus/templates/rules/alerts_service_prober.conf.yml
@@ -13,7 +13,7 @@ groups:
description: 'Probe {{ $labels.probe }} ({{ $labels.zone }}) is failing
for target {{ $labels.target }} (success ratio {{ $value }}). Check
https://service-prober.autistici.org/ for the error details.'
- runbook: '[[ alert_playbook_url ]]/ProbeFailure.md'
+ runbook: '[[ alert_playbook_url ]]/ProbeFailure'
- alert: ProbeFailure
expr: probe:probe_success:ratio{probeset="service"} < 0.5
@@ -26,5 +26,5 @@ groups:
description: 'Probe {{ $labels.probe }} ({{ $labels.zone }}) is failing
globally (success ratio {{ $value }}). Check
https://service-prober.autistici.org/ for the error details.'
- runbook: '[[ alert_playbook_url ]]/ProbeFailure.md'
+ runbook: '[[ alert_playbook_url ]]/ProbeFailure'