mostly fixes

[distro-setup] / filesystem / etc / prometheus / rules / iank.yml

diff --git a/filesystem/etc/prometheus/rules/iank.yml b/filesystem/etc/prometheus/rules/iank.yml
index 971e392fd63b55387089c47529060ebd91543d03..9838cac3ca219cbbf748da0ea366843056dbeb53 100644 (file)
--- a/filesystem/etc/prometheus/rules/iank.yml
+++ b/filesystem/etc/prometheus/rules/iank.yml
@@ -9,10 +9,9 @@ groups:
 - name: standard
   rules:
 
-## uncomment for testing an alert firing
+# ## uncomment for testing an alert firing
 #   - alert: test-alert4
 #     expr: vector(1)
-# #    expr: nonexistent_metric
 #     for: 0m
 #     labels:
 #       severity: day
@@ -107,30 +106,40 @@ groups:
     labels:
       severity: prod
 
+# 17 minutes: if we reboot causing 1 send to fail, thats 10 minutes. we
+# test this every 5 minutes, so thats 15 minutes at most.
   - alert: mailtest_check_vps
     expr: |-
-      time() - mailtest_check_last_usec{job="tlsnode"} >= 60 * 12
+      time() - mailtest_check_last_usec{job="tlsnode"} >= 60 * 17
     labels:
       severity: day
     annotations:
-      summary: '12 minutes down'
+      summary: '17 minutes down'
+
+  - alert: mailtest_check_unexpected_spamd_vps
+    expr: |-
+      mailtest_check_unexpected_spamd_results >= 1
+    labels:
+      severity: day
+    annotations:
+      summary: 'jr -u mailtest-check -e'
 
   - alert: mailtest_check_mailhost
     expr: |-
-      time() - max by (folder,from) (mailtest_check_last_usec{job="node"}) >= 60 * 12
+      time() - max by (folder,from) (mailtest_check_last_usec{job="node"}) >= 60 * 17
     labels:
       severity: day
     annotations:
-      summary: '12 minutes down'
+      summary: '17 minutes down'
 
-  # 42 mins: enough for a 30 min queue run plus 12
+  # 20 minutes. just allow for more due to prod alert.
   - alert: mailtest_check_gnu_mailhost
     expr: |-
-      time() - max by (folder,from) (mailtest_check_last_usec{folder="/m/md/l/testignore", from="iank@gnu.org"}) >= 60 * 42
+      time() - max by (folder,from) (mailtest_check_last_usec{folder="/m/md/l/testignore", from="iank@gnu.org"}) >= 60 * 20
     labels:
       severity: prod
     annotations:
-      summary: '42 minutes down'
+      summary: '20 minutes down'
 
 
   - alert: 1pmtest