services/monitoring: Setup
The goal is to create a monitoring setup where each server monitors itself when it comes failing systemd services, disk or RAM filling up, …. In addition each prometheus will monitor remote prometheus and alertmanager instances for signs of failure (e.g. being unreachable, errors in notification delivery, dropping alerts). A lot of metrics (especially histograms from prometheus or alertmanager) are being dropped before ingestion to disk save on space and memory. Depending on how many servers we may or may not have in the future this could probably use some kind of overhaul since we rightnow have n^2 monitoring peer relationships (not even speaking of possible duplicated alerts).
This commit is contained in:
Moritz 'e1mo' Fromm
parent
6e9c9a26cd
commit
3acc1865c0
4 changed files with 457 additions and 2 deletions
40
secrets/all/monitoring.htpasswd
Normal file
40
secrets/all/monitoring.htpasswd
Normal file
|
|
@ -0,0 +1,40 @@
|
||||||
|
{
|
||||||
|
"data": "ENC[AES256_GCM,data:/qJNDvVv7ap7CiCHbxtPEs35RepVZ/jSWeqaKtTMu4ePvZbpX+BDMDPztg==,iv:8p1gnTObH883P/h7QLtCYVbCWvhi6nKhkWNlvazy1CM=,tag:ha+5Mniu0On1TCAGTSCf3A==,type:str]",
|
||||||
|
"sops": {
|
||||||
|
"kms": null,
|
||||||
|
"gcp_kms": null,
|
||||||
|
"azure_kv": null,
|
||||||
|
"hc_vault": null,
|
||||||
|
"age": [
|
||||||
|
{
|
||||||
|
"recipient": "age14ysl953378r2vvy7ft3gwce9xp83pr6wypf5lgx2yjwx2lxra5qs6j8eqe",
|
||||||
|
"enc": "-----BEGIN AGE ENCRYPTED FILE-----\nYWdlLWVuY3J5cHRpb24ub3JnL3YxCi0+IFgyNTUxOSBJM21QS0VGNXNZV0NJSVU0\nYUxZaEpMQ0hQNkZPNFNzUjFOM3V2c1ZkaUY4CjdqY2tGaXFOL2liTlBxSitzSkx5\nQmMrTW5EYWZ3RWY0dEZ6NldVSkhBN28KLS0tIDEzamtzdEFQam5ZU3pJY1p3Lzh5\nY2wwdmpIWktWMEwyb09mVDh4Ujd2TUEKiIxFNC5fk38+GvBPIP0RygwA4uw49Uht\nfabstjYfQACI5Auc2IUL7nWPqVTT61obbogbGPfgRkOU/lnU7lhvtw==\n-----END AGE ENCRYPTED FILE-----\n"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"recipient": "age1w3wqxt5t00hjv43dcxlr5rjec5mvuzz9ajc8k04azq0gfx0ncgysu6mdmm",
|
||||||
|
"enc": "-----BEGIN AGE ENCRYPTED FILE-----\nYWdlLWVuY3J5cHRpb24ub3JnL3YxCi0+IFgyNTUxOSBIQzUraTdVYkg3THQzSXlS\neEtqSUNBMEIvSDZQNWVhNXdBaklPenk4aFhRCkU2UDE5UWZSdEFGV0tPYmxteUlh\nd2hZNXdXYSthK3dmalNrdi9MdE9ybm8KLS0tIFRTc3g2RFR3VWtyaDJBb2UvMWU3\nVkw5QW03djdyd2tCUndNNlpWcW1iUDgKF01T2t0pqk6O7gQh9xYn+KqtLXZRr3v3\nIjhMh46UOrm5OYSDqkKKFEU6Tz6PmHii9ClcC9ejClg+5lBaavlAqg==\n-----END AGE ENCRYPTED FILE-----\n"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"lastmodified": "2022-12-30T14:49:56Z",
|
||||||
|
"mac": "ENC[AES256_GCM,data:vCOGAJDFg1HH2+JwWqKWaSUAq0IS1voaTphRQHWPjXzgJ2eeDbXQQPEkp9Rk5Q5/zfPMdGgw8oR1OjWAv+KcaE/+t8PKPOd1B6a50VkFtgHYuHbunRai3RLMfexZMPfC7QvO5TLLNM4/OzrDsdVozPecj4LXpMrrC9rpacBCnWc=,iv:UtQvqNXzMC/kb0l0sG16PCZSIfacO7jEa88x5Y7eDS4=,tag:yXa2ohlBnh+kqJzstUYIWQ==,type:str]",
|
||||||
|
"pgp": [
|
||||||
|
{
|
||||||
|
"created_at": "2022-12-30T14:49:56Z",
|
||||||
|
"enc": "-----BEGIN PGP MESSAGE-----\n\nhF4DOnsoj685gdcSAQdA7WmJAs7/hgq93M3SphsDJv9m/oOLhf3ftXyngoj4MVkw\n4xKE1ZCN4LErd+/enCdGLeCkzxIFL7lPc1Q1BaCJZwAZjQIhwUFtK2ORWeXK+y65\n0l4Bon20voZRVNjqiU5XuzWTNUQOStiq30qf5MdCYTfyXQI7fpwYsIWoL3T3en41\nNsIHoDYmXgfoIGDHWmkFEW7YRy04/9wUvWnX67eQIEB9WPvX0s1nUnq0t/x7BVem\n=4Zp6\n-----END PGP MESSAGE-----\n",
|
||||||
|
"fp": "0x6D617FD0A85BAADA"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"created_at": "2022-12-30T14:49:56Z",
|
||||||
|
"enc": "-----BEGIN PGP MESSAGE-----\n\nhF4D6iFd6webPCUSAQdAk5UEGAJw64qu9tVkeaGGDkX4+9awdb3APedVb5EZPXIw\n46V4iFaF6U6mry2BqbPOzBZxE0AzQtmVR/H3SfPlrqJYak98lhYrRMfMjTs+7DAy\n0l4B4WIodnYaPhkhXuoq2gglCCyQ8FN3yUoDfYcIRgmmBtuwp+t3rPe4pJpDkf1t\nNyHhuKWRKtLaYc9YJMq6smpCICFFYV8RohVUmK6ucWGHJVC5wTX5+Cz/UZv+qpwB\n=tQa7\n-----END PGP MESSAGE-----\n",
|
||||||
|
"fp": "0xE0262A773B824745"
|
||||||
|
},
|
||||||
|
{
|
||||||
|
"created_at": "2022-12-30T14:49:56Z",
|
||||||
|
"enc": "-----BEGIN PGP MESSAGE-----\n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\n=a73T\n-----END PGP MESSAGE-----\n",
|
||||||
|
"fp": "B1480CFF9BBE8E2648A26A640B2E7C171E3AD6D7"
|
||||||
|
}
|
||||||
|
],
|
||||||
|
"unencrypted_suffix": "_unencrypted",
|
||||||
|
"version": "3.7.3"
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
@ -4,6 +4,8 @@ smtp:
|
||||||
user: ENC[AES256_GCM,data:cyxh+oTM7MdF31+umMplmA==,iv:maKuY4iXQCg10O8TTenBUl6tMJKt54AycfKQW6sheVw=,tag:4DAUiBBgO5H9O7DwLYZJtQ==,type:str]
|
user: ENC[AES256_GCM,data:cyxh+oTM7MdF31+umMplmA==,iv:maKuY4iXQCg10O8TTenBUl6tMJKt54AycfKQW6sheVw=,tag:4DAUiBBgO5H9O7DwLYZJtQ==,type:str]
|
||||||
pass: ENC[AES256_GCM,data:CkpgvLnQGBzShLkDaHtOHgZ9zwlsEJ+9KFBmHLgBol9Pcadww040clNM0WJgz8NfDIoYF8M=,iv:OIxuwC/XQOBBKdkRb6/5X8pNPLJNKAEI/WCYKhFScPw=,tag:CoAFk7n5AeSue18OXZvjKA==,type:str]
|
pass: ENC[AES256_GCM,data:CkpgvLnQGBzShLkDaHtOHgZ9zwlsEJ+9KFBmHLgBol9Pcadww040clNM0WJgz8NfDIoYF8M=,iv:OIxuwC/XQOBBKdkRb6/5X8pNPLJNKAEI/WCYKhFScPw=,tag:CoAFk7n5AeSue18OXZvjKA==,type:str]
|
||||||
host: ENC[AES256_GCM,data:psXf+lg9433jS1GxH6HQQQ==,iv:wyq/VgWLNs4RbzxNHf4Ave8ZDhQeNPpYXJIdjLLMlr4=,tag:pQ4Wmf+PzSBsHO3eigx91Q==,type:str]
|
host: ENC[AES256_GCM,data:psXf+lg9433jS1GxH6HQQQ==,iv:wyq/VgWLNs4RbzxNHf4Ave8ZDhQeNPpYXJIdjLLMlr4=,tag:pQ4Wmf+PzSBsHO3eigx91Q==,type:str]
|
||||||
|
alertmanager:
|
||||||
|
env: ENC[AES256_GCM,data:eYytFrXgsm4NCyS5F3oVNKsXwf5RL+zEFomMszP4fzb0HkLBk659uWZmqi7kIe8ptyUqqr6oAIV31alFo2TwpWuImvh4gbmaD0sfzesBmqhXyeA3cgi3EMGfMxsdLgUpHKGuJhgrRGLY9jDRI6VIwuLng1YmRw==,iv:YQYwC+Xef9I5IdUsBlTg3g7yRxWJe5Pi1LCur9g9/8w=,tag:U/A2lYu79urWjBFraQh16w==,type:str]
|
||||||
sops:
|
sops:
|
||||||
kms: []
|
kms: []
|
||||||
gcp_kms: []
|
gcp_kms: []
|
||||||
|
|
@ -28,8 +30,8 @@ sops:
|
||||||
UXludFJ5UDEvN2pHMEQweWIvN0NYaFEKB7lVKrsB3eX77iKWwFAXp7LVl+fPcGOl
|
UXludFJ5UDEvN2pHMEQweWIvN0NYaFEKB7lVKrsB3eX77iKWwFAXp7LVl+fPcGOl
|
||||||
8CkIBa/rkSWMe0xIetew60wIwx2ZVAv5TTDmKIZyQSTayOpbG3zcdg==
|
8CkIBa/rkSWMe0xIetew60wIwx2ZVAv5TTDmKIZyQSTayOpbG3zcdg==
|
||||||
-----END AGE ENCRYPTED FILE-----
|
-----END AGE ENCRYPTED FILE-----
|
||||||
lastmodified: "2022-12-24T21:21:02Z"
|
lastmodified: "2022-12-30T22:26:42Z"
|
||||||
mac: ENC[AES256_GCM,data:GbEO+lLUeswXR6Dc9kGlR9YJhXWh7Io6Z/qciolwkrzIwZAA6l6kaWCdw6OJ6gSNBGhWNye+VDdDV2Ow1Ix1hjnZdl3MUlY/LEEmzP1uUjPvNnEjZjxC8ebzNg3rgVqT6W6q9S7D9nPmAzJMgMlzypR/b1ZX5QuzodhotsleVNM=,iv:TJX940f70Vka+hsfDSvm7Jj0klFylRwB+9XbA9pFkRo=,tag:uaJi+5UI/bppLw+hOgyiDg==,type:str]
|
mac: ENC[AES256_GCM,data:KcJVntK3kAMH64ZHYoJmDCA7BroYJZhmMvtgbnjI44h5cwT9TyDOso2HO/x2z5XzTfPQfbW3tfhT1UmbaBHxBTkV7gJGU9fJ1xOrCRDoHuI9RTwGnCHKwQPLIWj56NMiC2CMf00pFpUlAPyDP/VHwRsc/laUPne2Da1MKZp/NOk=,iv:iHONd6xTjfVZNFELQVafAiljWYavu1DOZrS+d7pd8Zw=,tag:1Zh32h2OEtqKPoXy+8pypw==,type:str]
|
||||||
pgp:
|
pgp:
|
||||||
- created_at: "2022-12-23T19:08:38Z"
|
- created_at: "2022-12-23T19:08:38Z"
|
||||||
enc: |
|
enc: |
|
||||||
|
|
|
||||||
274
services/monitoring/default.nix
Normal file
274
services/monitoring/default.nix
Normal file
|
|
@ -0,0 +1,274 @@
|
||||||
|
{
|
||||||
|
config,
|
||||||
|
lib,
|
||||||
|
outputs,
|
||||||
|
...
|
||||||
|
}: let
|
||||||
|
inherit (lib) concatStringsSep mapAttrsToList hasAttrByPath getAttrFromPath filterAttrs substring singleton optionalString optional;
|
||||||
|
inherit (lib) escapeRegex;
|
||||||
|
inherit (config.networking) fqdn hostName;
|
||||||
|
|
||||||
|
# Basically a manual list of (legacy) hosts not yet migrated to NixOS
|
||||||
|
# but on which we'd like to have included in the monitoring.
|
||||||
|
externalTargets = let
|
||||||
|
host = hostName: {
|
||||||
|
_module.args.baseDomain = "chaos.jetzt";
|
||||||
|
config = {
|
||||||
|
networking = {
|
||||||
|
inherit hostName;
|
||||||
|
domain = "net.chaos.jetzt";
|
||||||
|
};
|
||||||
|
services.prometheus = {
|
||||||
|
enable = true;
|
||||||
|
port = 9090;
|
||||||
|
alertmanager = {
|
||||||
|
enable = true;
|
||||||
|
port = 9093;
|
||||||
|
};
|
||||||
|
};
|
||||||
|
};
|
||||||
|
};
|
||||||
|
in {
|
||||||
|
hopper = host "hopper";
|
||||||
|
lovelace = host "lovelace";
|
||||||
|
};
|
||||||
|
|
||||||
|
monDomain = "mon.${config.networking.domain}";
|
||||||
|
|
||||||
|
isMe = host: host.config.networking.fqdn == fqdn;
|
||||||
|
others = filterAttrs (_: !isMe) outputs.nixosConfigurations;
|
||||||
|
isDev = host: (substring 0 3 host._module.args.baseDomain) == "dev";
|
||||||
|
allHosts = outputs.nixosConfigurations // externalTargets;
|
||||||
|
/*
|
||||||
|
Right now we only have one non-dev host in our NixOS setup (the ansible hosts don't monitor the NixOS hosts).
|
||||||
|
That's why we currently add all hosts to our little monitoring "cluster". As soon as we have two or more production hosts,
|
||||||
|
the dev host can be taken out of the equation
|
||||||
|
*/
|
||||||
|
# allTargets = filterAttrs (_: c: (isMe c) || !(isDev c)) allHosts;
|
||||||
|
allTargets = allHosts;
|
||||||
|
|
||||||
|
# monFqdn = config: "${config.networking.hostName}.${monDomain}";
|
||||||
|
hasEnabled = servicePath: config: let
|
||||||
|
path = servicePath ++ ["enable"];
|
||||||
|
in
|
||||||
|
(hasAttrByPath path config) && (getAttrFromPath path config);
|
||||||
|
|
||||||
|
monTarget = servicePath: config: let
|
||||||
|
port = toString (getAttrFromPath (servicePath ++ ["port"]) config);
|
||||||
|
in "${config.networking.hostName}.${monDomain}:${port}";
|
||||||
|
|
||||||
|
dropMetrics = {wildcard ? true}: extraRegexen: let
|
||||||
|
dropRegexen = [ "go_" "promhttp_metric_handler_requests_" ] ++ extraRegexen;
|
||||||
|
in
|
||||||
|
singleton {
|
||||||
|
inherit (regex);
|
||||||
|
regex = "(${concatStringsSep "|" dropRegexen})${optionalString wildcard ".*"}";
|
||||||
|
source_labels = ["__name__"];
|
||||||
|
action = "drop";
|
||||||
|
};
|
||||||
|
|
||||||
|
relabelInstance = {
|
||||||
|
source_labels = ["__address__"];
|
||||||
|
regex = "(\\w+)\\.${escapeRegex monDomain}\\:\\d*";
|
||||||
|
target_label = "instance";
|
||||||
|
};
|
||||||
|
|
||||||
|
prometheusPath = ["services" "prometheus"];
|
||||||
|
alertmanagerPath = ["services" "prometheus" "alertmanager"];
|
||||||
|
targetAllHosts = servicePath:
|
||||||
|
mapAttrsToList
|
||||||
|
(_: config: monTarget servicePath config.config)
|
||||||
|
(filterAttrs (_: c: (hasEnabled servicePath c.config)) (outputs.nixosConfigurations // externalTargets));
|
||||||
|
in {
|
||||||
|
/*
|
||||||
|
Steps to edit the monitoring.htpasswd (aka. adding yourself / updating you password):
|
||||||
|
|
||||||
|
1. `sops -d secrets/all/monitoring.htpasswd > /tmp/monitoring.htpasswd`
|
||||||
|
2. Use `htpasswd` (from the `apacheHttpd` package) to your hearts content
|
||||||
|
3. `sops -e /tmp/monitoring.htpasswd > secrets/all/monitoring.htpasswd`
|
||||||
|
4. `rm /tmp/monitoring.htpasswd`
|
||||||
|
*/
|
||||||
|
sops.secrets = {
|
||||||
|
"monitoring.htpasswd" = {
|
||||||
|
format = "binary";
|
||||||
|
owner = config.services.nginx.user;
|
||||||
|
sopsFile = ../../secrets/all/monitoring.htpasswd;
|
||||||
|
};
|
||||||
|
"alertmanager/env" = {
|
||||||
|
format = "yaml";
|
||||||
|
sopsFile = ../../secrets/all/secrets.yaml;
|
||||||
|
};
|
||||||
|
};
|
||||||
|
|
||||||
|
services.nginx.virtualHosts."${fqdn}" = let
|
||||||
|
monitoring_htpasswd = config.sops.secrets."monitoring.htpasswd".path;
|
||||||
|
in {
|
||||||
|
enableACME = true;
|
||||||
|
forceSSL = true;
|
||||||
|
locations."/prometheus/" = {
|
||||||
|
basicAuthFile = monitoring_htpasswd;
|
||||||
|
proxyPass = "http://127.0.0.1:${builtins.toString config.services.prometheus.port}/";
|
||||||
|
};
|
||||||
|
locations."/alertmanager/" = {
|
||||||
|
basicAuthFile = monitoring_htpasswd;
|
||||||
|
proxyPass = "http://127.0.0.1:${builtins.toString config.services.prometheus.alertmanager.port}/";
|
||||||
|
};
|
||||||
|
};
|
||||||
|
|
||||||
|
services.prometheus.exporters.node = {
|
||||||
|
enable = true;
|
||||||
|
enabledCollectors = ["systemd"];
|
||||||
|
# They either don't apply to us or will provide us with metrics not usefull to us
|
||||||
|
disabledCollectors = [
|
||||||
|
"arp"
|
||||||
|
"bcache"
|
||||||
|
"bonding"
|
||||||
|
"btrfs"
|
||||||
|
"cpufreq"
|
||||||
|
"edac"
|
||||||
|
"entropy"
|
||||||
|
"infiniband"
|
||||||
|
"rapl"
|
||||||
|
"selinux"
|
||||||
|
"timex"
|
||||||
|
];
|
||||||
|
};
|
||||||
|
|
||||||
|
services.prometheus = {
|
||||||
|
enable = true;
|
||||||
|
webExternalUrl = "https://${fqdn}/prometheus/";
|
||||||
|
extraFlags = [
|
||||||
|
"--web.route-prefix=\"/\""
|
||||||
|
"--web.enable-admin-api"
|
||||||
|
];
|
||||||
|
ruleFiles = [
|
||||||
|
./rules.yaml
|
||||||
|
];
|
||||||
|
retentionTime = "30d";
|
||||||
|
|
||||||
|
alertmanagers = [{
|
||||||
|
static_configs = [{
|
||||||
|
targets = [(monTarget alertmanagerPath config)];
|
||||||
|
}];
|
||||||
|
}];
|
||||||
|
|
||||||
|
scrapeConfigs = [
|
||||||
|
{
|
||||||
|
job_name = "node";
|
||||||
|
static_configs = [{
|
||||||
|
targets = [
|
||||||
|
# Only scraping to own node-exporter
|
||||||
|
(monTarget ["services" "prometheus" "exporters" "node"] config)
|
||||||
|
];
|
||||||
|
}];
|
||||||
|
relabel_configs = [relabelInstance];
|
||||||
|
metric_relabel_configs = dropMetrics {} [];
|
||||||
|
}
|
||||||
|
{
|
||||||
|
job_name = "alertmanager";
|
||||||
|
static_configs = [{
|
||||||
|
targets = targetAllHosts alertmanagerPath;
|
||||||
|
}];
|
||||||
|
relabel_configs = [relabelInstance];
|
||||||
|
metric_relabel_configs = dropMetrics {} [
|
||||||
|
"alertmanager_http_(response_size_bytes|request_duration_seconds)_"
|
||||||
|
"alertmanager_notification_latency_seconds_"
|
||||||
|
"alertmanager_(nflog|cluster)_"
|
||||||
|
"alertmanager_silences_(query_duration_seconds|gc)_"
|
||||||
|
];
|
||||||
|
}
|
||||||
|
{
|
||||||
|
job_name = "prometheus";
|
||||||
|
static_configs = [{
|
||||||
|
targets = targetAllHosts prometheusPath;
|
||||||
|
}];
|
||||||
|
relabel_configs = [relabelInstance];
|
||||||
|
metric_relabel_configs = dropMetrics {} [
|
||||||
|
"prometheus_(sd|tsdb|target)_"
|
||||||
|
"prometheus_(engine_query|rule_evaluation)_duration_"
|
||||||
|
"prometheus_http_(response_size_bytes|request_duration_seconds)_"
|
||||||
|
"net_conntrack_dialer_conn_"
|
||||||
|
];
|
||||||
|
}
|
||||||
|
];
|
||||||
|
};
|
||||||
|
|
||||||
|
services.prometheus.alertmanager = {
|
||||||
|
enable = true;
|
||||||
|
extraFlags = ["--web.route-prefix=\"/\"" "--cluster.listen-address="];
|
||||||
|
webExternalUrl = "https://${fqdn}/alertmanager/";
|
||||||
|
environmentFile = config.sops.secrets."alertmanager/env".path;
|
||||||
|
|
||||||
|
configuration = {
|
||||||
|
global = {
|
||||||
|
smtp_from = "Chaos-Jetzt Monitoring (${hostName}) <monitoring-${hostName}@chaos.jetzt>";
|
||||||
|
smtp_smarthost = "\${SMTP_HOST}:587";
|
||||||
|
smtp_auth_username = "\${SMTP_USER}";
|
||||||
|
smtp_auth_password = "\${SMTP_PASS}";
|
||||||
|
smtp_hello = config.networking.fqdn;
|
||||||
|
};
|
||||||
|
|
||||||
|
receivers = [{
|
||||||
|
name = "mail";
|
||||||
|
email_configs = [
|
||||||
|
{ to = "jetzt+mon@e1mo.de";
|
||||||
|
send_resolved = true; }
|
||||||
|
{ to = "info@adb.sh";
|
||||||
|
send_resolved = true; }
|
||||||
|
];
|
||||||
|
}];
|
||||||
|
|
||||||
|
route = {
|
||||||
|
receiver = "mail";
|
||||||
|
repeat_interval = "16h";
|
||||||
|
group_wait = "1m";
|
||||||
|
group_by = ["alertname" "instance"];
|
||||||
|
routes = [
|
||||||
|
{
|
||||||
|
match.severiy = "critical";
|
||||||
|
receiver = "mail";
|
||||||
|
repeat_interval = "6h";
|
||||||
|
}
|
||||||
|
{
|
||||||
|
match.severiy = "error";
|
||||||
|
receiver = "mail";
|
||||||
|
repeat_interval = "16h";
|
||||||
|
}
|
||||||
|
{
|
||||||
|
match.severiy = "warn";
|
||||||
|
receiver = "mail";
|
||||||
|
repeat_interval = "28h";
|
||||||
|
}
|
||||||
|
{
|
||||||
|
match.severiy = "info";
|
||||||
|
receiver = "mail";
|
||||||
|
repeat_interval = "56h";
|
||||||
|
}
|
||||||
|
];
|
||||||
|
};
|
||||||
|
|
||||||
|
inhibit_rules = [
|
||||||
|
{
|
||||||
|
target_matchers = ["alertname = ReducedAvailableMemory"];
|
||||||
|
source_matchers = ["alertname =~ (Very)LowAvailableMemory"];
|
||||||
|
equal = ["instance"];
|
||||||
|
}
|
||||||
|
{
|
||||||
|
target_matchers = ["alertname = LowAvailableMemory"];
|
||||||
|
source_matchers = ["alertname = VeryLowAvailableMemory"];
|
||||||
|
equal = ["instance"];
|
||||||
|
}
|
||||||
|
{
|
||||||
|
target_matchers = ["alertname = ElevatedLoad"];
|
||||||
|
source_matchers = ["alertname =~ (Very)HighLoad"];
|
||||||
|
equal = ["instance"];
|
||||||
|
}
|
||||||
|
{
|
||||||
|
target_matchers = ["alertname = HighLoad"];
|
||||||
|
source_matchers = ["alertname = VeryHighLoad"];
|
||||||
|
equal = ["instance"];
|
||||||
|
}
|
||||||
|
];
|
||||||
|
};
|
||||||
|
};
|
||||||
|
}
|
||||||
139
services/monitoring/rules.yaml
Normal file
139
services/monitoring/rules.yaml
Normal file
|
|
@ -0,0 +1,139 @@
|
||||||
|
---
|
||||||
|
groups:
|
||||||
|
- name: General system
|
||||||
|
rules:
|
||||||
|
- record: node_memory_MemAvailable_percentage
|
||||||
|
expr: (node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100
|
||||||
|
- record: node_cpu_count
|
||||||
|
expr: count without(cpu, mode) (node_cpu_seconds_total{mode="idle"})
|
||||||
|
|
||||||
|
- alert: InstanceDown
|
||||||
|
expr: up == 0
|
||||||
|
for: 5m
|
||||||
|
labels:
|
||||||
|
serverity: critical
|
||||||
|
annotations:
|
||||||
|
summary: "{{ $labels.instance }} down"
|
||||||
|
|
||||||
|
- alert: SystemdServiceFailed
|
||||||
|
expr: node_systemd_unit_state{state="failed"} > 0
|
||||||
|
for: 5m
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "{{ $labels.name }} on {{ $labels.instance }} crashed"
|
||||||
|
|
||||||
|
- alert: OomKill
|
||||||
|
expr: increase(node_vmstat_oom_kill[5m]) > 0 # Scrape interval of 1 minute
|
||||||
|
for: 0m
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "OOM kill on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: VeryLowAvailableMemory
|
||||||
|
# Less than 3% available or less than 100KB available memory
|
||||||
|
expr: ((node_memory_MemAvailable_percentage < 3) or (node_memory_MemFree_bytes < (1024 * 100)))
|
||||||
|
for: 2m
|
||||||
|
labels:
|
||||||
|
serverity: critical
|
||||||
|
annotations:
|
||||||
|
summary: "Very low free memory on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: LowAvailableMemory
|
||||||
|
expr: node_memory_MemAvailable_percentage < 10
|
||||||
|
for: 5m
|
||||||
|
labels:
|
||||||
|
serverity: warn
|
||||||
|
annotations:
|
||||||
|
summary: "Low free memory on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: ReducedAvailableMemory
|
||||||
|
expr: node_memory_MemAvailable_percentage < 20
|
||||||
|
# Detect constant high memory usage as a potential sign that the host could maybe need some more memory
|
||||||
|
for: 1h
|
||||||
|
labels:
|
||||||
|
serverity: info
|
||||||
|
annotations:
|
||||||
|
summary: "Reduced available memory on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: HighMemoryPressure
|
||||||
|
# For at least 5 seconds over the last 2 Minutes, no progress could be made
|
||||||
|
# due to memory congestion
|
||||||
|
expr: increase(node_pressure_memory_stalled_seconds_total[2m]) >= 5
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "Memory preassure on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: VeryHighLoad
|
||||||
|
expr: ((node_load1 / node_cpu_count) > 1) or ((node_load5 / node_cpu_count) > 0.9)
|
||||||
|
for: 5m
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "Very High load on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: HighLoad
|
||||||
|
expr: ((node_load1 / node_cpu_count) > 0.9) or ((node_load5 / node_cpu_count) > 0.8)
|
||||||
|
for: 10m
|
||||||
|
labels:
|
||||||
|
serverity: warn
|
||||||
|
annotations:
|
||||||
|
summary: "High load on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: ElevatedLoad
|
||||||
|
# Detecting a long-term increased load
|
||||||
|
expr: (node_load15 / node_cpu_count) > 0.7
|
||||||
|
for: 15m
|
||||||
|
labels:
|
||||||
|
serverity: info
|
||||||
|
annotations:
|
||||||
|
summary: "Elevated load15 on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: LowFreeDiskSpace
|
||||||
|
expr: node_filesystem_readonly == 0 and ON (instance, device, mountpoint) (node_filesystem_avail_bytes * 100) / node_filesystem_size_bytes < 15
|
||||||
|
for: 10m
|
||||||
|
labels:
|
||||||
|
serverity: info
|
||||||
|
annotations:
|
||||||
|
summary: "{{ $labels.mountpoint }} on {{ $labels.instance }} has less than 15% space left"
|
||||||
|
|
||||||
|
- alert: VeryLowFreeDiskSpace
|
||||||
|
expr: node_filesystem_readonly == 0 and ON (instance, device, mountpoint) (node_filesystem_avail_bytes * 100) / node_filesystem_size_bytes < 5
|
||||||
|
for: 3m
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "{{ $labels.mountpoint }} on {{ $labels.instance }} has less than 5% space left"
|
||||||
|
|
||||||
|
- name: Prometheus Alertmanager
|
||||||
|
rules:
|
||||||
|
- alert: AlertmanagerMissing
|
||||||
|
expr: prometheus_notifications_alertmanagers_discovered < 1
|
||||||
|
for: 5m
|
||||||
|
labels:
|
||||||
|
serverity: warn
|
||||||
|
annotations:
|
||||||
|
summary: "No connected alertmanager on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: AlertmanagerMissing
|
||||||
|
expr: increase(prometheus_notifications_dropped_total[5m]) > 0
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "Dropped alert notifications on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: FailingAlertmanagerNotifications
|
||||||
|
expr: increase(alertmanager_notifications_failed_total[5m]) > 0
|
||||||
|
labels:
|
||||||
|
serverity: error
|
||||||
|
annotations:
|
||||||
|
summary: "Failing notifications via {{ $labels.integration }} on {{ $labels.instance }}"
|
||||||
|
|
||||||
|
- alert: FailingRuleEvaluation
|
||||||
|
expr: sum by (instance) (increase(prometheus_rule_evaluation_failures_total[5m])) > 0
|
||||||
|
labels:
|
||||||
|
serverity: warn
|
||||||
|
annotations:
|
||||||
|
summary: "Failing rule evaluations on {{ $labels.instance }}"
|
||||||
Loading…
Reference in a new issue