Fix statuscake alerts

[chef.git] / cookbooks / prometheus / recipes / server.rb

diff --git a/cookbooks/prometheus/recipes/server.rb b/cookbooks/prometheus/recipes/server.rb
index 20275ab7ac5e5e52f1656ce4d313e8dff51596ee..ffc951e065880d4cf877f961c7fd2f5ed5ad1103 100644 (file)
--- a/cookbooks/prometheus/recipes/server.rb
+++ b/cookbooks/prometheus/recipes/server.rb
@@ -18,7 +18,8 @@
 #
 
 include_recipe "apache"
-include_recipe "apt"
+include_recipe "apt::grafana"
+include_recipe "apt::timescaledb"
 include_recipe "networking"
 include_recipe "timescaledb"
 
@@ -34,19 +35,22 @@ end
 
 prometheus_exporter "fastly_healthcheck" do
   port 9696
+  scrape_interval "1m"
   environment "FASTLY_API_TOKEN" => tokens["fastly"]
 end
 
 prometheus_exporter "statuscake" do
   port 9595
+  scrape_interval "5m"
+  scrape_timeout "2m"
   environment "STATUSCAKE_APIKEY" => tokens["statuscake"]
 end
 
 cache_dir = Chef::Config[:file_cache_path]
 
-prometheus_version = "2.31.1"
-alertmanager_version = "0.23.0"
-karma_version = "0.93"
+prometheus_version = "2.41.0"
+alertmanager_version = "0.24.0"
+karma_version = "0.111"
 
 directory "/opt/prometheus-server" do
   owner "root"
@@ -69,7 +73,7 @@ archive_file "#{cache_dir}/prometheus.linux-amd64.tar.gz" do
   strip_components 1
   owner "root"
   group "root"
-  subscribes :extract, "remote_file[#{cache_dir}/prometheus.linux-amd64.tar.gz]"
+  subscribes :extract, "remote_file[#{cache_dir}/prometheus.linux-amd64.tar.gz]", :immediately
 end
 
 remote_file "#{cache_dir}/alertmanager.linux-amd64.tar.gz" do
@@ -87,7 +91,7 @@ archive_file "#{cache_dir}/alertmanager.linux-amd64.tar.gz" do
   strip_components 1
   owner "root"
   group "root"
-  subscribes :extract, "remote_file[#{cache_dir}/alertmanager.linux-amd64.tar.gz]"
+  subscribes :extract, "remote_file[#{cache_dir}/alertmanager.linux-amd64.tar.gz]", :immediately
 end
 
 remote_file "#{cache_dir}/karma-linux-amd64.tar.gz" do
@@ -104,20 +108,29 @@ archive_file "#{cache_dir}/karma-linux-amd64.tar.gz" do
   overwrite true
   owner "root"
   group "root"
-  subscribes :extract, "remote_file[#{cache_dir}/karma-linux-amd64.tar.gz]"
+  subscribes :extract, "remote_file[#{cache_dir}/karma-linux-amd64.tar.gz]", :immediately
 end
 
-promscale_version = "0.11.0"
+promscale_version = "0.17.0"
+promscale_extension_version = "0.8.0-1"
 
 database_version = node[:timescaledb][:database_version]
 database_cluster = "#{database_version}/main"
 
 package %W[
   prometheus
-  prometheus-alertmanager
   promscale-extension-postgresql-#{database_version}
 ]
 
+package "promscale-extension-postgresql-#{database_version}" do
+  version promscale_extension_version
+end
+
+apt_preference "promscale-extension-postgresql" do
+  pin "version #{promscale_extension_version}"
+  pin_priority "1100"
+end
+
 postgresql_user "prometheus" do
   cluster database_cluster
   superuser true
@@ -154,10 +167,8 @@ systemd_service "promscale" do
   user "prometheus"
   exec_start "/opt/promscale/bin/promscale --db.uri postgresql:///promscale?host=/run/postgresql&port=5432 --db.connections-max 400"
   limit_nofile 16384
-  private_tmp true
-  protect_system "strict"
-  protect_home true
-  no_new_privileges true
+  sandbox :enable_network => true
+  restrict_address_families "AF_UNIX"
 end
 
 if node[:prometheus][:promscale]
@@ -201,11 +212,15 @@ search(:node, "recipes:prometheus\\:\\:default").sort_by(&:name).each do |client
       name = exporter[:name]
       address = exporter[:address]
       sni = exporter[:sni]
+      scrape_interval = exporter[:scrape_interval]
+      scrape_timeout = exporter[:scrape_timeout]
       metric_relabel = exporter[:metric_relabel] || []
     else
       name = key
       address = exporter
       sni = nil
+      scrape_interval = nil
+      scrape_timeout = nil
       metric_relabel = []
     end
 
@@ -214,6 +229,8 @@ search(:node, "recipes:prometheus\\:\\:default").sort_by(&:name).each do |client
       :address => address,
       :sni => sni,
       :instance => client.name.split(".").first,
+      :scrape_interval => scrape_interval,
+      :scrape_timeout => scrape_timeout,
       :metric_relabel => metric_relabel
     }
   end
@@ -285,13 +302,21 @@ service "prometheus" do
   subscribes :restart, "archive_file[#{cache_dir}/prometheus.linux-amd64.tar.gz]"
 end
 
-systemd_service "prometheus-alertmanager-executable" do
-  service "prometheus-alertmanager"
-  dropin "executable"
+systemd_service "prometheus-alertmanager" do
+  description "Prometheus alert manager"
+  type "simple"
+  user "prometheus"
   exec_start "/opt/prometheus-server/alertmanager/alertmanager --config.file=/etc/prometheus/alertmanager.yml --storage.path=/var/lib/prometheus/alertmanager --web.external-url=https://prometheus.openstreetmap.org/alertmanager"
+  exec_reload "/bin/kill -HUP $MAINPID"
+  timeout_stop_sec 20
+  restart "on-failure"
   notifies :restart, "service[prometheus-alertmanager]"
 end
 
+link "/usr/local/bin/promtool" do
+  to "/opt/prometheus-server/prometheus/promtool"
+end
+
 template "/etc/prometheus/alertmanager.yml" do
   source "alertmanager.yml.erb"
   owner "root"
@@ -299,19 +324,36 @@ template "/etc/prometheus/alertmanager.yml" do
   mode "644"
 end
 
+directory "/var/lib/prometheus/alertmanager" do
+  owner "prometheus"
+  group "prometheus"
+  mode "755"
+end
+
 service "prometheus-alertmanager" do
   action [:enable, :start]
   subscribes :reload, "template[/etc/prometheus/alertmanager.yml]"
+  subscribes :restart, "systemd_service[prometheus-alertmanager]"
   subscribes :restart, "archive_file[#{cache_dir}/alertmanager.linux-amd64.tar.gz]"
 end
 
-template "/etc/prometheus/amtool.yml" do
+directory "/etc/amtool" do
+  owner "root"
+  group "root"
+  mode "755"
+end
+
+template "/etc/amtool/config.yml" do
   source "amtool.yml.erb"
   owner "root"
   group "root"
   mode "644"
 end
 
+link "/usr/local/bin/amtool" do
+  to "/opt/prometheus-server/alertmanager/amtool"
+end
+
 template "/etc/prometheus/karma.yml" do
   source "karma.yml.erb"
   owner "root"
@@ -323,18 +365,15 @@ systemd_service "prometheus-karma" do
   description "Alert dashboard for Prometheus Alertmanager"
   user "prometheus"
   exec_start "/opt/prometheus-server/karma/karma-linux-amd64 --config.file=/etc/prometheus/karma.yml"
-  private_tmp true
-  private_devices true
-  protect_system "full"
-  protect_home true
-  no_new_privileges true
+  sandbox :enable_network => true
   restart "on-failure"
 end
 
 service "prometheus-karma" do
   action [:enable, :start]
-  subscribes :reload, "template[/etc/prometheus/karma.yml]"
+  subscribes :restart, "template[/etc/prometheus/karma.yml]"
   subscribes :restart, "archive_file[#{cache_dir}/karma-linux-amd64.tar.gz]"
+  subscribes :restart, "systemd_service[prometheus-karma]"
 end
 
 package "grafana-enterprise"