X-Git-Url: https://git.openstreetmap.org./chef.git/blobdiff_plain/93b82f6239bf810ecdb96241c7680628f5a7ebc9..c0fc233e4d310ddf4593838f21e2eee39c34d871:/cookbooks/prometheus/recipes/server.rb?ds=inline diff --git a/cookbooks/prometheus/recipes/server.rb b/cookbooks/prometheus/recipes/server.rb index 675a206ec..405eadebe 100644 --- a/cookbooks/prometheus/recipes/server.rb +++ b/cookbooks/prometheus/recipes/server.rb @@ -28,8 +28,7 @@ admins = data_bag_item("apache", "admins") prometheus_exporter "fastly" do port 8080 - listen_switch "endpoint" - listen_type "url" + listen_switch "listen" environment "FASTLY_API_TOKEN" => tokens["fastly"] end @@ -37,6 +36,7 @@ cache_dir = Chef::Config[:file_cache_path] prometheus_version = "2.31.1" alertmanager_version = "0.23.0" +karma_version = "0.93" directory "/opt/prometheus-server" do owner "root" @@ -80,16 +80,34 @@ archive_file "#{cache_dir}/alertmanager.linux-amd64.tar.gz" do subscribes :extract, "remote_file[#{cache_dir}/alertmanager.linux-amd64.tar.gz]" end -package %w[ - prometheus - prometheus-alertmanager -] +remote_file "#{cache_dir}/karma-linux-amd64.tar.gz" do + source "https://github.com/prymitive/karma/releases/download/v#{karma_version}/karma-linux-amd64.tar.gz" + owner "root" + group "root" + mode "644" + backup false +end + +archive_file "#{cache_dir}/karma-linux-amd64.tar.gz" do + action :nothing + destination "/opt/prometheus-server/karma" + overwrite true + owner "root" + group "root" + subscribes :extract, "remote_file[#{cache_dir}/karma-linux-amd64.tar.gz]" +end -promscale_version = "0.6.2" +promscale_version = "0.11.0" database_version = node[:timescaledb][:database_version] database_cluster = "#{database_version}/main" +package %W[ + prometheus + prometheus-alertmanager + promscale-extension-postgresql-#{database_version} +] + postgresql_user "prometheus" do cluster database_cluster superuser true @@ -106,13 +124,6 @@ directory "/opt/promscale" do mode "755" end -cookbook_file "/usr/lib/postgresql/#{database_version}/lib/promscale.so" do - source "postgresql-#{database_version}-promscale.so" - owner "root" - group "root" - mode "644" -end - directory "/opt/promscale/bin" do owner "root" group "root" @@ -131,8 +142,7 @@ systemd_service "promscale" do description "Promscale Connector" type "simple" user "prometheus" - exec_start "/opt/promscale/bin/promscale --db-uri postgresql:///promscale?host=/run/postgresql&port=5432 --db-connections-max 400" - # exec_start lazy { "/opt/promscale/bin/promscale --db-host /run/postgresql --db-port #{node[:postgresql][:clusters][database_cluster][:port]} --db-user prometheus --db-name promscale --db-max-connections 400" } + 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" @@ -140,41 +150,16 @@ systemd_service "promscale" do no_new_privileges true end -systemd_service "promscale-maintenance" do - description "Promscale Maintenance" - type "simple" - user "prometheus" - exec_start "/usr/bin/psql --command='CALL prom_api.execute_maintenance()' promscale" - private_tmp true - protect_system "strict" - protect_home true - no_new_privileges true -end - -systemd_timer "promscale-maintenance" do - description "Promscale Maintenace" - on_active_sec 1800 - on_unit_inactive_sec 1800 -end - if node[:prometheus][:promscale] service "promscale" do action [:enable, :start] subscribes :restart, "remote_file[/opt/promscale/bin/promscale]" subscribes :restart, "systemd_service[promscale]" end - - service "promscale-maintenance.timer" do - action [:enable, :start] - end else service "promscale" do action [:disable, :stop] end - - service "promscale-maintenance.timer" do - action [:disable, :stop] - end end search(:node, "roles:gateway") do |gateway| @@ -205,16 +190,19 @@ search(:node, "recipes:prometheus\\:\\:default").sort_by(&:name).each do |client if exporter.is_a?(Hash) name = exporter[:name] address = exporter[:address] + sni = exporter[:sni] metric_relabel = exporter[:metric_relabel] || [] else name = key address = exporter + sni = nil metric_relabel = [] end jobs[name] ||= [] jobs[name] << { :address => address, + :sni => sni, :instance => client.name.split(".").first, :metric_relabel => metric_relabel } @@ -283,6 +271,7 @@ service "prometheus" do action [:enable, :start] subscribes :reload, "template[/etc/prometheus/prometheus.yml]" subscribes :reload, "template[/etc/prometheus/alert_rules.yml]" + subscribes :restart, "archive_file[#{cache_dir}/prometheus.linux-amd64.tar.gz]" end systemd_service "prometheus-alertmanager-executable" do @@ -302,6 +291,7 @@ end service "prometheus-alertmanager" do action [:enable, :start] subscribes :reload, "template[/etc/prometheus/alertmanager.yml]" + subscribes :restart, "archive_file[#{cache_dir}/alertmanager.linux-amd64.tar.gz]" end template "/etc/prometheus/amtool.yml" do @@ -311,6 +301,31 @@ template "/etc/prometheus/amtool.yml" do mode "644" end +template "/etc/prometheus/karma.yml" do + source "karma.yml.erb" + owner "root" + group "root" + mode "644" +end + +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 + restart "on-failure" +end + +service "prometheus-karma" do + action [:enable, :start] + subscribes :reload, "template[/etc/prometheus/karma.yml]" + subscribes :restart, "archive_file[#{cache_dir}/karma-linux-amd64.tar.gz]" +end + package "grafana-enterprise" template "/etc/grafana/grafana.ini" do