From 24fccea832af75f8416ba9e3e2df3f298d510632 Mon Sep 17 00:00:00 2001
From: Peter Sabaini <peter.sabaini@canonical.com>
Date: Wed, 17 Jan 2024 16:50:37 +0100
Subject: [PATCH] Add alerting rules for RGW multisite deployments

Add default prometheus alerting rules for RadosGW multisite deployments based
on the built-in Ceph RGW multisite metrics.

Note that the included prometheus_alerts.yml.default rule file
is included for reference only. The ceph-mon charm will utilize the
resource file from https://charmhub.io/ceph-mon/resources/alert-rules
for deployment so that operators can easily customize these rules.

Change-Id: I5a12162d73686963132a952bddd85ec205964de4
---
 .../prometheus_alerts.yml.default             | 47 +++++++++++++++++++
 1 file changed, 47 insertions(+)

diff --git a/files/prometheus_alert_rules/prometheus_alerts.yml.default b/files/prometheus_alert_rules/prometheus_alerts.yml.default
index a544d41e..e0914e49 100644
--- a/files/prometheus_alert_rules/prometheus_alerts.yml.default
+++ b/files/prometheus_alert_rules/prometheus_alerts.yml.default
@@ -633,3 +633,50 @@ groups:
           oid: "1.3.6.1.4.1.50495.1.2.1.1.2"
           severity: "critical"
           type: "ceph_default"
+  - name: "rgwmultisite"
+    rules:
+      - alert: "CephRGWMultisiteFetchError"
+        annotations:
+          description: "Unsuccessful Object Replications from source zone threshold has been exceeded. The threshold is defined as 2 errors per 15min"
+          summary: "Unsuccessful Object Replications from Source Zone Threshold Exceeded"
+        expr: "increase(ceph_data_sync_from_zone_fetch_errors[15m]) > 2"
+        for: "5m"
+        labels:
+          severity: "warning"
+          type: "ceph_default"
+      - alert: "CephRGWMultisitePollError"
+        annotations:
+          description: "Unsuccessful Object Replications from Source Zone Threshold Exceeded. The threshold is defined as 2 errors per 15min"
+          summary: "Unsuccessful Object Replications from Source Zone Threshold Exceeded"
+        expr: "increase(ceph_data_sync_from_zone_poll_errors[15m]) > 2"
+        for: "5m"
+        labels:
+          severity: "warning"
+          type: "ceph_default"
+      - alert: "CephRGWMultisiteFetchErrorCritical"
+        annotations:
+          description: "Critical: Unsuccessful Object Replications from source zone threshold has been exceeded. The threshold is defined as 50 errors per 15min"
+          summary: "Critical: Unsuccessful Object Replications from Source Zone Threshold Exceeded"
+        expr: "increase(ceph_data_sync_from_zone_fetch_errors[15m]) > 50"
+        for: "5m"
+        labels:
+          severity: "critical"
+          type: "ceph_default"
+      - alert: "CephRGWMultisitePollErrorCritical"
+        annotations:
+          description: "Critical: Unsuccessful Object Replications from source zone threshold has been exceeded. The threshold is defined as 50 errors per 15min"
+          summary: "Critical: Unsuccessful Object Replications from Source Zone Threshold Exceeded"
+        expr: "increase(ceph_data_sync_from_zone_poll_errors[15m]) > 50"
+        for: "5m"
+        labels:
+          severity: "critical"
+          type: "ceph_default"
+      - alert: "CephRGWMultisitePollLatency"
+        annotations:
+          description: "Latency for poll request threshold exceeded. The threshold is defined as 600s latency per 15min"
+          summary: "Poll Request Latency Threshold Exceeded"
+        expr: "increase(ceph_data_sync_from_zone_poll_latency_sum[15m]) > 600"
+        for: "5m"
+        labels:
+          severity: "warning"
+          type: "ceph_default"