Cloonar/nixos
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

feat: restructure Grafana configuration, migrate alert rules to new format and add VictoriaMetrics datasource

Browse Source
This commit is contained in:
Dominik Polakovics Polakovics
2025-05-31 09:27:25 +02:00
parent 17a3602d3c
commit 8b5fb0861d
14 changed files with 384 additions and 166 deletions
Download Patch File Download Diff File Expand all files Collapse all files

66
hosts/web-arm/modules/grafana/alerting/cpu_usage.nix Normal file
View File

@@ -0,0 +1,66 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
name = "CPUUsageAlerts";
folder = "System Alerts";
interval = "1m";
rules = [
{
uid = "high-cpu-usage-alert-uid";
title = "HighCPUUsage";
condition = "D"; # Condition is now D
data = [
# Query A: Calculate CPU usage percentage
{
refId = "A";
datasourceUid = "vm-datasource-uid";
queryType = "prometheus";
relativeTimeRange = { from = 60; to = 0; }; # Query over the last minute
model = {
# Calculate average CPU usage over 1m, grouped by instance and job
expr = ''(1 - avg by (instance, job) (rate(node_cpu_seconds_total{mode="idle"}[1m]))) * 100'';
legendFormat = "CPU usage on {{instance}} ({{job}})";
instant = false; # This is a range query
};
}
# Expression C: Reduce Query A to its last value, preserving labels
{
refId = "C";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
# Expression D: Apply math condition to the reduced values from C
{
refId = "D";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$C > 90"; # Alert if CPU usage from C is > 90%
};
}
];
for = "5m"; # Duration the condition must be met
noDataState = "NoData";
execErrState = "Error";
annotations = {
summary = "High CPU usage on {{ $labels.instance }}";
description = ''CPU usage on {{ $labels.instance }} (job: {{ $labels.job }}) has been above 90% for more than 5 minutes. Current value: {{ if $values.C }}{{ $values.C | humanizePercentage }}{{ else }}N/A{{ end }}%.'';
};
labels = {
severity = "warning";
category = "performance";
};
}
];
}
];
}

85
hosts/web-arm/modules/grafana/alerting/disk_usage.nix Normal file
View File

@@ -0,0 +1,85 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
# orgId = 1; # Defaults to 1 for provisioned rules
name = "DiskUsageAlerts"; # Name of the rule group
folder = "System Alerts"; # The folder these rules belong to in Grafana UI
interval = "1m"; # How often to evaluate rules in this group
rules = [
{
uid = "high-disk-usage-alert-uid"; # Optional: provide a stable UID for the rule itself
title = "HighDiskUsage"; # Name of the alert rule (was 'alert' in vmalert)
# Condition for the alert to fire. 'C' refers to the refId of the threshold expression.
condition = "D"; # Condition is now D
# Removed rule-level relativeTimeRange
# Data queries and expressions
data = [
# Query A: Calculate disk usage percentage
{
refId = "A";
datasourceUid = "vm-datasource-uid"; # UID of the VictoriaMetrics datasource
queryType = "prometheus"; # Explicitly set, though often inferred
relativeTimeRange = { from = 60; to = 0; }; # Query-level, integer seconds
model = {
expr = ''
(
node_filesystem_size_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""} - node_filesystem_avail_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""}
) / (node_filesystem_size_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""} > 0) * 100
and node_filesystem_size_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""}
and node_filesystem_avail_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""}
'';
legendFormat = "{{mountpoint}} on {{instance}}"; # Example legend
instant = false; # For range queries, default is false
};
}
# Expression C: Reduce Query A to its last value, preserving labels
{
refId = "C";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
# Expression D: Apply math condition to the reduced values from C
{
refId = "D";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$C > 85"; # Check if the last value from each series in C is > 85
};
}
];
for = "15m"; # Duration the condition must be met (same as vmalert)
# How to handle states where data is missing or query errors
noDataState = "NoData"; # Options: NoData, Alerting, OK
execErrState = "Error"; # Options: Error, Alerting, OK
annotations = {
summary = "High disk usage on {{ $labels.instance }} at {{ $labels.mountpoint }}";
description = ''
Disk usage on {{ $labels.instance }} for mount point {{ $labels.mountpoint }}
(fstype: {{ $labels.fstype }}) has been above 85% for more than 15 minutes.
Current value: {{ if $values.C }}{{ $values.C | humanizePercentage }}{{ else }}N/A{{ end }}%.
''; # Using $values.C as it's the input to the math condition D
};
labels = {
severity = "warning";
category = "capacity";
# Grafana automatically adds labels from the query result (instance, mountpoint, etc.)
# and labels from the rule group/folder.
};
# isPaused = false; # Default is not paused
}
];
}
];
}

62
hosts/web-arm/modules/grafana/alerting/host_down.nix Normal file
View File

@@ -0,0 +1,62 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
name = "HostStatusAlerts";
folder = "System Alerts";
interval = "1m";
rules = [
{
uid = "host-down-alert-uid";
title = "HostDown";
condition = "C";
data = [
{
refId = "A";
datasourceUid = "vm-datasource-uid";
queryType = "prometheus";
relativeTimeRange = { from = 60; to = 0; }; # Query over the last minute
model = {
expr = ''up'';
legendFormat = "{{instance}} ({{job}})";
instant = false; # Changed from true, as relativeTimeRange is used
};
}
{ # New Expression B: Reduce Query A
refId = "B";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
{ # Modified Expression C: Math condition based on B
refId = "C";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$B == 0"; # Check if the last value from B is 0
};
}
];
for = "2m";
noDataState = "Alerting";
execErrState = "Error";
annotations = {
summary = "Host {{ $labels.instance }} is down";
description = ''Host {{ $labels.instance }} (job: {{ $labels.job }}) has been down for more than 2 minutes.'';
};
labels = {
severity = "critical";
category = "availability";
};
}
];
}
];
}

71
hosts/web-arm/modules/grafana/alerting/inode_usage.nix Normal file
View File

@@ -0,0 +1,71 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
name = "InodeUsageAlerts";
folder = "System Alerts";
interval = "1m";
rules = [
{
uid = "high-inode-usage-alert-uid";
title = "HighInodeUsage";
condition = "D"; # Condition is now D
data = [
# Query A: Calculate inode usage percentage
{
refId = "A";
datasourceUid = "vm-datasource-uid";
queryType = "prometheus";
relativeTimeRange = { from = 60; to = 0; };
model = {
expr = ''
(
node_filesystem_files{fstype!~"tmpfs|rootfs",mountpoint!=""} - node_filesystem_files_free{fstype!~"tmpfs|rootfs",mountpoint!=""}
) / (node_filesystem_files{fstype!~"tmpfs|rootfs",mountpoint!=""} > 0) * 100
and node_filesystem_files{fstype!~"tmpfs|rootfs",mountpoint!=""}
and node_filesystem_files_free{fstype!~"tmpfs|rootfs",mountpoint!=""}
'';
legendFormat = "{{mountpoint}} on {{instance}}";
instant = false;
};
}
# Expression C: Reduce Query A to its last value, preserving labels
{
refId = "C";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
# Expression D: Apply math condition to the reduced values from C
{
refId = "D";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$C > 80"; # Alert if inode usage from C is > 80%
};
}
];
for = "30m"; # Duration the condition must be met
noDataState = "NoData";
execErrState = "Error";
annotations = {
summary = "High inode usage on {{ $labels.instance }} at {{ $labels.mountpoint }}";
description = ''Inode usage on {{ $labels.instance }} for mount point {{ $labels.mountpoint }} (fstype: {{ $labels.fstype }}) has been above 80% for more than 30 minutes. Current value: {{ if $values.C }}{{ $values.C | humanizePercentage }}{{ else }}N/A{{ end }}%.'';
};
labels = {
severity = "warning";
category = "capacity";
};
}
];
}
];
}

69
hosts/web-arm/modules/grafana/alerting/ram_usage.nix Normal file
View File

@@ -0,0 +1,69 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.alerting.rules.settings.groups = [
{
name = "RAMUsageAlerts";
folder = "System Alerts";
interval = "1m";
rules = [
{
uid = "high-ram-usage-alert-uid";
title = "HighRAMUsage";
condition = "D"; # Condition is now D
data = [
# Query A: Calculate RAM usage percentage
{
refId = "A";
datasourceUid = "vm-datasource-uid";
queryType = "prometheus";
relativeTimeRange = { from = 60; to = 0; };
model = {
expr = ''
(1 - node_memory_MemAvailable_bytes / (node_memory_MemTotal_bytes > 0)) * 100
and node_memory_MemAvailable_bytes
and node_memory_MemTotal_bytes
'';
legendFormat = "RAM usage on {{instance}} ({{job}})";
instant = false;
};
}
# Expression C: Reduce Query A to its last value, preserving labels
{
refId = "C";
datasourceUid = "__expr__";
model = {
type = "reduce";
expression = "A"; # Input is Query A
reducer = "last"; # Get the last value of each series in A
};
}
# Expression D: Apply math condition to the reduced values from C
{
refId = "D";
datasourceUid = "__expr__";
model = {
type = "math";
expression = "$C > 90"; # Alert if RAM usage from C is > 90%
};
}
];
for = "10m"; # Duration the condition must be met
noDataState = "NoData";
execErrState = "Error";
annotations = {
summary = "High RAM usage on {{ $labels.instance }}";
description = ''RAM usage on {{ $labels.instance }} (job: {{ $labels.job }}) has been above 90% for more than 10 minutes. Current value: {{ if $values.C }}{{ $values.C | humanizePercentage }}{{ else }}N/A{{ end }}%.'';
};
labels = {
severity = "warning";
category = "performance";
};
}
];
}
];
}

18
hosts/web-arm/modules/grafana/datasources/victoriametrics.nix Normal file
View File

@@ -0,0 +1,18 @@
{ lib, pkgs, config, ... }:
{
services.grafana.provision.datasources.settings.datasources = [
{
name = "VictoriaMetrics";
uid = "vm-datasource-uid"; # Stable UID for referencing in alerts
type = "prometheus";
url = "http://localhost:8428"; # URL of VictoriaMetrics
access = "proxy"; # Grafana proxies requests
isDefault = true; # Optional: make this the default datasource
jsonData = {
# timeInterval = "30s"; # Optional: Scrape interval if different from Grafana's default
# httpMethod = "POST"; # Optional: if VictoriaMetrics prefers POST for queries
};
editable = false; # Recommended for provisioned datasources
}
];
}

12
hosts/web-arm/modules/grafana.nix → hosts/web-arm/modules/grafana/default.nix
View File

@@ -27,6 +27,16 @@ let
};
in
{
imports = [
./alerting/disk_usage.nix
./alerting/cpu_usage.nix
./alerting/host_down.nix
./alerting/inode_usage.nix
./alerting/ram_usage.nix
# ... other rule files can be added here ...
./datasources/victoriametrics.nix
];
systemd.services.grafana.script = lib.mkBefore ''
export GF_AUTH_GENERIC_OAUTH_CLIENT_SECRET=$(cat /run/secrets/grafana-oauth-secret)
export PUSHOVER_API_TOKEN=$(cat /run/secrets/pushover-api-token)
@@ -89,6 +99,7 @@ in
};
provision = {
alerting = {
rules.settings.groups = lib.mkMerge []; # Allows rule groups to be merged
contactPoints = {
settings = {
apiVersion = 1; # As per Grafana provisioning API
@@ -158,6 +169,7 @@ in
};
};
};
datasources.settings.datasources = lib.mkMerge []; # Allows datasources to be merged
};
};

38
hosts/web-arm/modules/vmalert/default.nix
View File

@@ -1,38 +0,0 @@
{ config, pkgs, lib, ... }:
{
imports = [
./rules/cpu_usage.nix
./rules/disk_usage.nix
./rules/host_down.nix
./rules/inode_usage.nix
./rules/ram_usage.nix
];
# Standard vmalert service configuration
services.vmalert = {
enable = true;
settings = {
"datasource.url" = "http://localhost:8428"; # VictoriaMetrics address
"notifier.url" = [ "http://localhost:3001/api/alertmanager/grafana/api/v2/alerts" ]; # Must be a list of strings
};
# 'rules' is now set by the mkMerge block above.
};
# Override the User and Group for the systemd service managed by the official vmalert module.
systemd.services.vmalert = {
serviceConfig = {
User = "victoriametrics";
Group = "victoriametrics";
};
};
# Ensure the user/group itself exists on the system.
users.users.victoriametrics = lib.mkIf (config.services.victoriametrics.enable || config.services.vmalert.enable) {
isSystemUser = true;
group = "victoriametrics"; # Primary group for the user
home = "/var/lib/victoriametrics"; # Standard home for VictoriaMetrics components
};
users.groups.victoriametrics = lib.mkIf (config.services.victoriametrics.enable || config.services.vmalert.enable) {
# Ensures the group exists.
};
}

26
hosts/web-arm/modules/vmalert/rules/cpu_usage.nix
View File

@@ -1,26 +0,0 @@
{ lib, pkgs, config, ... }: # Standard module arguments
{
# This module contributes its rule group to a list that will be
# collected and processed by the main vmalert module.
services.vmalert.rules.groups = [
{
name = "CPUUsageAlerts";
# interval = "60s"; # Optional: group-level interval
rules = [ # This MUST be a list of rule attribute sets
{
alert = "HighCPUUsage";
expr = "(1 - avg by (instance, job) (rate(node_cpu_seconds_total{mode=\"idle\"}[1m]))) * 100 > 90";
for = "5m";
labels = {
severity = "warning";
category = "performance";
};
annotations = {
summary = "High CPU usage on {{ $labels.instance }}";
description = "CPU usage on {{ $labels.instance }} (job: {{ $labels.job }}) has been above 90% for more than 5 minutes. Current value: {{ $value | printf \"%.2f\" }}%.";
};
}
];
}
];
}

27
hosts/web-arm/modules/vmalert/rules/disk_usage.nix
View File

@@ -1,27 +0,0 @@
{ lib, pkgs, config, ... }: # Standard module arguments
{
services.vmalert.rules.groups = [
{
name = "DiskUsageAlerts";
rules = [
{
alert = "HighDiskUsage";
expr = ''
(
node_filesystem_size_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""} - node_filesystem_avail_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""}
) / node_filesystem_size_bytes{fstype!~"tmpfs|rootfs",mountpoint!=""} * 100 > 85
'';
for = "15m";
labels = {
severity = "warning";
category = "capacity";
};
annotations = {
summary = "High disk usage on {{ $labels.instance }} at {{ $labels.mountpoint }}";
description = "Disk usage on {{ $labels.instance }} for mount point {{ $labels.mountpoint }} (fstype: {{ $labels.fstype }}) has been above 85% for more than 15 minutes. Current value: {{ $value | printf \"%.2f\" }}%.";
};
}
];
}
];
}

23
hosts/web-arm/modules/vmalert/rules/host_down.nix
View File

@@ -1,23 +0,0 @@
{ lib, pkgs, config, ... }: # Standard module arguments
{
services.vmalert.rules.groups = [
{
name = "HostStatusAlerts";
rules = [
{
alert = "HostDown";
expr = "up == 0";
for = "2m";
labels = {
severity = "critical";
category = "availability";
};
annotations = {
summary = "Host {{ $labels.instance }} is down";
description = "Host {{ $labels.instance }} (job: {{ $labels.job }}) has been down for more than 2 minutes.";
};
}
];
}
];
}

27
hosts/web-arm/modules/vmalert/rules/inode_usage.nix
View File

@@ -1,27 +0,0 @@
{ lib, pkgs, config, ... }: # Standard module arguments
{
services.vmalert.rules.groups = [
{
name = "InodeUsageAlerts";
rules = [
{
alert = "HighInodeUsage";
expr = ''
(
node_filesystem_files{fstype!~"tmpfs|rootfs",mountpoint!=""} - node_filesystem_files_free{fstype!~"tmpfs|rootfs",mountpoint!=""}
) / node_filesystem_files{fstype!~"tmpfs|rootfs",mountpoint!=""} * 100 > 80
'';
for = "30m";
labels = {
severity = "warning";
category = "capacity";
};
annotations = {
summary = "High inode usage on {{ $labels.instance }} at {{ $labels.mountpoint }}";
description = "Inode usage on {{ $labels.instance }} for mount point {{ $labels.mountpoint }} (fstype: {{ $labels.fstype }}) has been above 80% for more than 30 minutes. Current value: {{ $value | printf \"%.2f\" }}%.";
};
}
];
}
];
}

23
hosts/web-arm/modules/vmalert/rules/ram_usage.nix
View File

@@ -1,23 +0,0 @@
{ lib, pkgs, config, ... }: # Standard module arguments
{
services.vmalert.rules.groups = [
{
name = "RAMUsageAlerts";
rules = [
{
alert = "HighRAMUsage";
expr = "(1 - node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100 > 90";
for = "10m";
labels = {
severity = "warning";
category = "performance";
};
annotations = {
summary = "High RAM usage on {{ $labels.instance }}";
description = "RAM usage on {{ $labels.instance }} (job: {{ $labels.job }}) has been above 90% for more than 10 minutes. Current value: {{ $value | printf \"%.2f\" }}%.";
};
}
];
}
];
}