nixpkgs/nixos/tests/elk.nix

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

277 lines
10 KiB
Nix
Raw Normal View History

2023-05-20 04:10:21 +02:00
# To run the test on the unfree ELK use the following command:
# cd path/to/nixpkgs
# NIXPKGS_ALLOW_UNFREE=1 nix-build -A nixosTests.elk.unfree.ELK-7
{ system ? builtins.currentSystem,
config ? {},
pkgs ? import ../.. { inherit system config; },
}:
let
inherit (pkgs) lib;
esUrl = "http://localhost:9200";
mkElkTest = name : elk :
import ./make-test-python.nix ({
inherit name;
meta = with pkgs.lib.maintainers; {
maintainers = [ eelco offline basvandijk ];
};
nodes = {
one =
{ pkgs, lib, ... }: {
# Not giving the machine at least 2060MB results in elasticsearch failing with the following error:
#
# OpenJDK 64-Bit Server VM warning:
# INFO: os::commit_memory(0x0000000085330000, 2060255232, 0)
# failed; error='Cannot allocate memory' (errno=12)
#
# There is insufficient memory for the Java Runtime Environment to continue.
# Native memory allocation (mmap) failed to map 2060255232 bytes for committing reserved memory.
#
# When setting this to 2500 I got "Kernel panic - not syncing: Out of
# memory: compulsory panic_on_oom is enabled" so lets give it even a
# bit more room:
virtualisation.memorySize = 3000;
# For querying JSON objects returned from elasticsearch and kibana.
environment.systemPackages = [ pkgs.jq ];
services = {
journalbeat = {
enable = elk ? journalbeat;
package = elk.journalbeat;
2020-01-08 10:02:26 +01:00
extraConfig = pkgs.lib.mkOptionDefault (''
logging:
to_syslog: true
level: warning
metrics.enabled: false
output.elasticsearch:
hosts: [ "127.0.0.1:9200" ]
journalbeat.inputs:
- paths: []
seek: cursor
'');
};
filebeat = {
enable = elk ? filebeat;
package = elk.filebeat;
inputs.journald.id = "everything";
inputs.log = {
enabled = true;
paths = [
"/var/lib/filebeat/test"
];
};
settings = {
logging.level = "info";
};
};
2021-03-19 17:16:10 +01:00
metricbeat = {
enable = true;
package = elk.metricbeat;
modules.system = {
metricsets = ["cpu" "load" "memory" "network" "process" "process_summary" "uptime" "socket_summary"];
enabled = true;
period = "5s";
processes = [".*"];
cpu.metrics = ["percentages" "normalized_percentages"];
core.metrics = ["percentages"];
};
settings = {
output.elasticsearch = {
hosts = ["127.0.0.1:9200"];
};
};
};
logstash = {
enable = true;
package = elk.logstash;
inputConfig = ''
exec { command => "echo -n flowers" interval => 1 type => "test" }
exec { command => "echo -n dragons" interval => 1 type => "test" }
'';
filterConfig = ''
if [message] =~ /dragons/ {
drop {}
}
'';
outputConfig = ''
file {
path => "/tmp/logstash.out"
codec => line { format => "%{message}" }
}
elasticsearch {
hosts => [ "${esUrl}" ]
}
'';
};
elasticsearch = {
enable = true;
package = elk.elasticsearch;
};
2018-08-21 10:39:25 +02:00
elasticsearch-curator = {
enable = elk ? elasticsearch-curator;
2018-08-21 10:39:25 +02:00
actionYAML = ''
---
actions:
1:
action: delete_indices
description: >-
Delete indices older than 1 second (based on index name), for logstash-
2018-08-21 10:39:25 +02:00
prefixed indices. Ignore the error if the filter does not result in an
actionable list of indices (ignore_empty_list) and exit cleanly.
options:
allow_ilm_indices: true
2018-08-21 10:39:25 +02:00
ignore_empty_list: True
disable_action: False
filters:
- filtertype: pattern
kind: prefix
value: logstash-
- filtertype: age
source: name
direction: older
timestring: '%Y.%m.%d'
unit: seconds
2018-08-21 10:39:25 +02:00
unit_count: 1
'';
};
};
};
};
2021-03-22 14:53:05 +01:00
passthru.elkPackages = elk;
testScript =
let
valueObject = lib.optionalString (lib.versionAtLeast elk.elasticsearch.version "7") ".value";
in ''
2020-01-08 10:02:26 +01:00
import json
def expect_hits(message):
dictionary = {"query": {"match": {"message": message}}}
return (
"curl --silent --show-error --fail-with-body '${esUrl}/_search' "
+ "-H 'Content-Type: application/json' "
+ "-d '{}' ".format(json.dumps(dictionary))
+ " | tee /dev/console"
+ " | jq -es 'if . == [] then null else .[] | .hits.total${valueObject} > 0 end'"
)
def expect_no_hits(message):
2020-01-08 10:02:26 +01:00
dictionary = {"query": {"match": {"message": message}}}
return (
"curl --silent --show-error --fail-with-body '${esUrl}/_search' "
2020-01-08 10:02:26 +01:00
+ "-H 'Content-Type: application/json' "
+ "-d '{}' ".format(json.dumps(dictionary))
+ " | tee /dev/console"
+ " | jq -es 'if . == [] then null else .[] | .hits.total${valueObject} == 0 end'"
2020-01-08 10:02:26 +01:00
)
2021-03-19 17:16:10 +01:00
def has_metricbeat():
dictionary = {"query": {"match": {"event.dataset": {"query": "system.cpu"}}}}
return (
"curl --silent --show-error --fail-with-body '${esUrl}/_search' "
2021-03-19 17:16:10 +01:00
+ "-H 'Content-Type: application/json' "
+ "-d '{}' ".format(json.dumps(dictionary))
+ " | tee /dev/console"
+ " | jq -es 'if . == [] then null else .[] | .hits.total${valueObject} > 0 end'"
2021-03-19 17:16:10 +01:00
)
2020-01-08 10:02:26 +01:00
start_all()
2020-01-08 10:02:26 +01:00
one.wait_for_unit("elasticsearch.service")
one.wait_for_open_port(9200)
# Continue as long as the status is not "red". The status is probably
# "yellow" instead of "green" because we are using a single elasticsearch
# node which elasticsearch considers risky.
#
# TODO: extend this test with multiple elasticsearch nodes
# and see if the status turns "green".
2020-01-08 10:02:26 +01:00
one.wait_until_succeeds(
"curl --silent --show-error --fail-with-body '${esUrl}/_cluster/health'"
+ " | jq -es 'if . == [] then null else .[] | .status != \"red\" end'"
2020-01-08 10:02:26 +01:00
)
with subtest("Perform some simple logstash tests"):
one.wait_for_unit("logstash.service")
one.wait_until_succeeds("cat /tmp/logstash.out | grep flowers")
one.wait_until_succeeds("cat /tmp/logstash.out | grep -v dragons")
2021-03-19 17:16:10 +01:00
with subtest("Metricbeat is running"):
one.wait_for_unit("metricbeat.service")
with subtest("Metricbeat metrics arrive in elasticsearch"):
one.wait_until_succeeds(has_metricbeat())
2021-03-19 17:16:10 +01:00
2020-01-08 10:02:26 +01:00
with subtest("Logstash messages arive in elasticsearch"):
one.wait_until_succeeds(expect_hits("flowers"))
one.wait_until_succeeds(expect_no_hits("dragons"))
2020-01-08 10:02:26 +01:00
'' + lib.optionalString (elk ? journalbeat) ''
2020-01-08 10:02:26 +01:00
with subtest(
"A message logged to the journal is ingested by elasticsearch via journalbeat"
):
one.wait_for_unit("journalbeat.service")
one.execute("echo 'Supercalifragilisticexpialidocious' | systemd-cat")
one.wait_until_succeeds(
expect_hits("Supercalifragilisticexpialidocious")
2020-01-08 10:02:26 +01:00
)
'' + lib.optionalString (elk ? filebeat) ''
with subtest(
"A message logged to the journal is ingested by elasticsearch via filebeat"
):
one.wait_for_unit("filebeat.service")
one.execute("echo 'Superdupercalifragilisticexpialidocious' | systemd-cat")
one.wait_until_succeeds(
expect_hits("Superdupercalifragilisticexpialidocious")
)
one.execute(
"echo 'SuperdupercalifragilisticexpialidociousIndeed' >> /var/lib/filebeat/test"
)
one.wait_until_succeeds(
expect_hits("SuperdupercalifragilisticexpialidociousIndeed")
)
'' + lib.optionalString (elk ? elasticsearch-curator) ''
2020-01-08 10:02:26 +01:00
with subtest("Elasticsearch-curator works"):
one.systemctl("stop logstash")
one.systemctl("start elasticsearch-curator")
one.wait_until_succeeds(
'! curl --silent --show-error --fail-with-body "${esUrl}/_cat/indices" | grep logstash | grep ^'
2020-01-08 10:02:26 +01:00
)
'';
}) { inherit pkgs system; };
in {
# We currently only package upstream binaries.
# Feel free to package an SSPL licensed source-based package!
2021-11-03 20:35:11 +01:00
# ELK-7 = mkElkTest "elk-7-oss" {
# name = "elk-7";
# elasticsearch = pkgs.elasticsearch7-oss;
# logstash = pkgs.logstash7-oss;
# filebeat = pkgs.filebeat7;
# metricbeat = pkgs.metricbeat7;
# };
unfree = lib.dontRecurseIntoAttrs {
ELK-7 = mkElkTest "elk-7" {
elasticsearch = pkgs.elasticsearch7;
logstash = pkgs.logstash7;
filebeat = pkgs.filebeat7;
metricbeat = pkgs.metricbeat7;
};
};
}