summaryrefslogtreecommitdiffstats
path: root/krebs/2configs/shack/prometheus/alert-rules.nix
diff options
context:
space:
mode:
authormakefu <github@syntax-fehler.de>2020-12-16 16:10:08 +0100
committermakefu <github@syntax-fehler.de>2020-12-16 16:10:08 +0100
commitdb80207267dd750d6e5fce0a4c15961aa324627b (patch)
tree8bfb1eafaeb723d945067cbfb2ddafdb4c371023 /krebs/2configs/shack/prometheus/alert-rules.nix
parentdeb3343d1eff747f37d08730d187574677540bf7 (diff)
ma anon-sftp: init
Diffstat (limited to 'krebs/2configs/shack/prometheus/alert-rules.nix')
-rw-r--r--krebs/2configs/shack/prometheus/alert-rules.nix7
1 files changed, 4 insertions, 3 deletions
diff --git a/krebs/2configs/shack/prometheus/alert-rules.nix b/krebs/2configs/shack/prometheus/alert-rules.nix
index 1c2d0b1ad..12c691466 100644
--- a/krebs/2configs/shack/prometheus/alert-rules.nix
+++ b/krebs/2configs/shack/prometheus/alert-rules.nix
@@ -14,7 +14,7 @@ in {
labels.severity = "warning";
annotations.summary = "{{ $labels.alias }} root disk full";
annotations.url = "http://grafana.shack/d/hb7fSE0Zz/shack-system-dashboard?orgId=1&var-job=node&var-hostname=All&var-node=wolf.shack:9100&var-device=All&var-maxmount=%2F&var-show_hostname=wolf";
- annotations.description = ''The root disk of {{ $labels.alias }} has {{ $value | printf "%.2f" }}% free disk space (Threshold at ${disk_free_threshold}%).A vast number of shackspace services will stop working. CI for deploying new configuration will also seize working. Log in to the system and run `nix-collect-garbage -d` and clean up the shack share folder in `/home/share` .If this does not help you can check `du -hs /var/ | sort -h`, run `docker system prune` or if you are really desperate run `du -hs / | sort -h` and go through the folders recursively until you've found something to delete'';
+ annotations.description = ''The root disk of {{ $labels.alias }} has {{ $value | printf "%.2f" }}% free disk space (Threshold at ${disk_free_threshold}%). CI for deploying new configuration will seize working. Log in to the system and run `nix-collect-garbage -d` and clean up the shack share folder in `/home/share` .If this does not help you can check `du -hs /var/ | sort -h`, run `docker system prune` or if you are really desperate run `du -hs / | sort -h` and go through the folders recursively until you've found something to delete'';
}
{
alert = "RootPartitionFull";
@@ -25,14 +25,15 @@ in {
annotations.url = "http://grafana.shack/d/hb7fSE0Zz/shack-system-dashboard?orgId=1&var-job=node&var-hostname=All&var-node=wolf.shack:9100&var-device=All&var-maxmount=%2F&var-show_hostname=puyak";
annotations.description = ''The root disk of {{ $labels.alias }} has {{ $value | printf "%.2f" }}% free disk space (Threshold at ${disk_free_threshold}%).Prometheus will not be able to create new alerts and CI for deploying new configuration will also seize working. Log in to the system and run `nix-collect-garbage -d` and if this does not help you can check `du -hs /var/ | sort -h`, run `docker system prune` or if you are really desperate run `du -hs / | sort -h` and go through the folders recursively until you've found something to delete'';
}
+ # wolf.shack is not worth supervising anymore
{
alert = "HostDown";
- expr = ''up{alias="wolf.shack"} == 0'';
+ expr = ''up{alias="infra01.shack"} == 0'';
for = "5m";
labels.severity = "page";
annotations.summary = "Instance {{ $labels.alias }} down for 5 minutes";
annotations.url = "http://grafana.shack/d/hb7fSE0Zz/shack-system-dashboard?orgId=1&var-job=node&var-hostname=All&var-node=wolf.shack:9100&var-device=All&var-maxmount=%2F&var-show_hostname=wolf";
- annotations.description = ''Host {{ $labels.alias }} went down and has not been reconnected after 5 minutes. This is probably bad news, try to restart the host via naproxen ( http://naproxen.shack:8006 ). Wolf being down means that CI,glados automation, light management and a couple of other services will not work anymore.'';
+ annotations.description = ''Host {{ $labels.alias }} went down and has not been reconnected after 5 minutes. This is probably bad news, as the machine runs one of the DNS servers and the power broadcast proxy which is used to be able to turn off the light via puyak as well as the shutdown listener.'';
}
];
}