keywords | title | nb_path | layout |
---|---|---|---|
fastai |
Checks in InlfuxDB |
_notebooks/2020-11-12-influxdb-check.ipynb |
notebook |
{% raw %}
You can use InfluxDB to monitor your data and generate notifications reporting its status.
- Check: generate statuses based on a queries expressing a conditino. It stores them in the
_monitoring
bucket (understatuses
measurement) - Notification rule: check the
statuses
measurement and, based on its own conditions, triggers the notification endpoint. It also stores notifications in the_monitoring
bucket (undernotifications
measurement) - Notification endpoint:
The elements of a check:
- Query: Defines what to monitor.
- Configuration: defines check interval, messages, and status.
Check types
- Threshold check: monitors a value against (high/low) thresholds and assigns one of the following values: CRIT, WARN, INFO, OK.
- Deadman check: monitors if a series or group responds within a given amount of time.
- Select the Alerts menu in the navigation pane.
- Click on Create
- Select the check type.
By default, you will find yourself in the Define Query screen.
- Select the bucket, the measurement, the field, and the tag sets.
- For threshold check: Select an aggregate function (e.g. mean, median, std, etc.).
You can preview the results of the query by clicking on Submit.
After these steps, select the Configure Check screen.
- Define the check schedule interval (this updates the aggregate function interval).
- Filter the check based on tags (Optional).
- Edit the status message template (Optional).
- For threshold check: define the thresholds and rules that will trigger the specific statuses. You can also set them with the mouse by moving their respective labels on the preview screen.
- For deadman check: specify the duration to check for, the status the lack of series or groups should trigger, and time after which it should stop checking the series / group.
Finally, name the check at the top of the screen.
Click the green checkmark to save the check.