Skip to content

Latest commit

 

History

History
232 lines (158 loc) · 6.89 KB

README.md

File metadata and controls

232 lines (158 loc) · 6.89 KB
Raw

sample plugins for Kibana

Warning: these are all in various states of disrepair. But usually easy to get working again in Kibana with tweaks.

The easiest way to use these is to symlink them into your kibana/plugins directory. Run the ln-plugins.sh command to link the plugins locally into a peer kibana clone directory. Ie, this git repo clone should be in the same directory as the kibana repo clone you want it installed in.

el_constipation - detect event loop blockages

This plugin will log to the console when the event log has been blocked for a configured threshold.

One http entrypoint is provided to force the event loop to be blocked for a specified amount of time. The following curl invocation will block the event loop for 5 seconds (5000 milliseconds).

export KBN_URLBASE=https://elastic:changeme@localhost:5601
curl -k -X POST -H "kbn-xsrf: X" $KBN_URLBASE/_dev/eat_cpu?millis=5000

v8_profiling - V8 CPU Profiling and Heap Snapshots

This plugin adds the following http entrypoints:

/_dev/cpu_profile?duration=<number>

Captures a CPU profile for duration seconds.

duration defaults to 5.

/_dev/heap_snapshot

Captures a heap snapshot.

example invocations

curl -k "https://elastic:changeme@localhost:5601/_dev/cpu_profile?duration=10" > my.cpuprofile

curl -k "https://elastic:changeme@localhost:5601/_dev/heap_snapshot" > my.heapsnapshot

Note the file extensions .cpuprofile and .heapsnapshot are required when loading the files into the V8's Chrome Dev Tools via the URL chrome://inspect/.

You can also use those URLs in the browser. When used, they will download the relevant file to your "Downloads" directory with the relevant file extension and a time-stamp based name.

index_threshold_graph - display index threshold alert data

See the index_threshold_graph README for more info on this plugin.

alert_type_examples - Alert Type Examples

This plugin adds the following alert types:

  • example.always-firing
  • example.essql
  • example.fizz-buzz
  • example.heartbeat

To play with the alerts in an existing kibana dev environment, you will need to do the following:

  • install kbn-actions

  • install jq

  • for Kibana 7.5 or lower, add the following lines to config kibana.dev.yml

    xpack.actions.enabled: true
xpack.alerting.enabled: true

  • in one terminal, start es via:

    yarn es snapshot --ssl

  • in another terminal, start kibana via:

    yarn start --no-base-path --ssl

  • open another terminal to run kbn-action and kbn-alert commands, running the following first to set the endpoint for those commands:

    export KBN_URLBASE=https://elastic:changeme@localhost:5601

You will probably want to create a server-log action that you can use with these alerts; do that via:

export ACTION_ID=`kbn-action create .server-log 'server-log' '{}' '{}' | jq -r '.id'`

Note that stores the action id in the env var ACTION_ID, used in subsequent commands.

Test it via:

kbn-action execute $ACTION_ID '{level: info, message: hallo}'

You should see a line hallo logged to the Kibana server log.

After creating a bunch of alerts, you'll probably want to delete them, and perhaps actions you've created. Use the following commands to delete all existing alerts and actions:

kbn-alert  ls | jq -r '.data | .[] | .id' | xargs -L 1 kbn-alert  delete
kbn-action ls | jq -r '.data | .[] | .id' | xargs -L 1 kbn-action delete

example.always-firing

An alert that always fires on it's interval.

This alert always fires the default group, with the following properties in it's context:

  • date - date/time the alert fired, in ISO format
  • count - number of times the alert function has run

example invocation

kbn-alert create example.always-firing "always firing example" 1s \
  '{}' \
  "[
    {group:default id:'${ACTION_ID}' params:{level: info, message:
      'alert example.always-firing, date: {{context.date}}; count: {{context.count}}'}
    }
  ]"

example.essql

An alert that runs an es sql query on an interval, and fires when hits are found.

documentation TBD

example invocation

kbn-alert create example.essql "essql example" 30s \
  '{query: "SELECT host.id as instanceId, \"@timestamp\" as timeStamp, system.cpu.system.pct as cpu FROM \"metricbeat-*\" WHERE system.cpu.system.pct > 1.5 AND timeStamp > NOW() - INTERVAL 30 SECONDS"}' \
  "[
    {group: hits, id: '$ACTION_ID', params: {level: info, message:
      'host {{instanceId}} at cpu {{cpu}} at {{timeStamp}}'
    }}
  ]"

example.fizz-buzz

An alert that fires groups fizz, buzz, and fizz-buzz on various intervals.

example invocation

kbn-alert create example.fizz-buzz "fizz-buzz example" 1s \
  '{}' \
  "[ 
    {group: fizz,        id: '$ACTION_ID', params: {level: info, message: 'fizz {{context.count}}'}}
    {group: buzz,        id: '$ACTION_ID', params: {level: info, message: 'buzz {{context.count}}'}}
    {group: 'fizz-buzz', id: '$ACTION_ID', params: {level: info, message: 'fizz-buzz {{context.count}}'}}
   ]"

example.heartbeat

An alert that works off of uptime heartbeat documents.

You can use es-hb-sim as a simulator to feed data to an elasticsearch index, in the shape this alert is expecting.

This alert takes two parameters:

  • index - the name of the index to search for uptime heartbeat data
  • window - the number of minutes worth of most recent data to access

There are four groups for this alert:

  • up - the monitor has transitioned from a non-up state to up
  • down - the monitor is down
  • flapping - the monitor has reported being up and down
  • noData - not enough data to make an analysis

The context passed to the action contains:

  • startedAt - ISO string of date the alert execution started
  • alertId - id of the alert
  • spaceId - spaceId of the alert (string)
  • namespace - namespace of the alert (string or undefined for 'default' spaceId)
  • name - name of the alert
  • tags - tags of the alert
  • createdBy - userid that created the alert
  • updatedBy - userid that last updated the alert
  • monitor - the name of the monitor affected by this alert

example invocation

(note window: 0.17 is 0.17 minutes == 10 seconds, useful to test with)

kbn-alert create example.heartbeat 'example heartbeat' 1s \
  '{index: "monitor-sample" window: 0.17}' \
  "[
    {group:up       id:'$ACTION_ID' params:{level:info, message: 'up: {{context.monitor}}'}}
    {group:down     id:'$ACTION_ID' params:{level:info, message: 'down: {{context.monitor}}'}}
    {group:flapping id:'$ACTION_ID' params:{level:info, message: 'flapping: {{context.monitor}}'}}
    {group:noData   id:'$ACTION_ID' params:{level:info, message: 'noData: {{context.monitor}}'}}
   ]"