Skip to main content
Version: Develop

Disk Alerts

Setup Disk Usage Alerts

Each service instance node comes with the a9s Parachute component. This component monitors ephemeral and persistent disk usage. If one of these disk usages reaches the default configured threshold of 80%, the a9s Parachute stops all processes on that node. Once the processes gets stopped, a9s Parachute logs the following messages to its log directory:

Limit reached for: persistent disk

or

Limit reached for: ephemeral disk

How to restart stopped instances

To restart a stopped instance, the user can use the cf update-service command to set the max_disk_threshold custom parameter. All values between 0 and 100 are allowed. Setting max_disk_threshold to 0 disables disk usage alert. Updating the instance does two things, at first the limit is set to the new value and the instance will be restarted. To set the value to 90% do the following:

$ cf update-service <SERVICE_NAME> -c '{"max_disk_threshold": 90}'

The update will take some minutes until the service is restarted.

IMPORTANT: This is only a temporary solution. The nodes will be stopped when the threshold is reached again. For a long term solution the service instance must be updated to a bigger plan.

Set the Disk Threshold Service Wide

You can configure the max_disk_threshold not only for a service instance. You can configure it for every service plan.

To do so, you will need to extend the default_placeholder_values in the service configuration.

Note:

  • If you want to configure the max_disk_threshold only for one specific plan (for example mariadb-single-small), you will need to adjust only this specific plan.
  • If you want to configure the max_disk_threshold for the complete data service offering, you will need to adjust every service plan in the configuration file.

Important

If you specify the max_disk_threshold on the service plan level, you are no longer able to overwrite it with cf update-service.

Example

You want to configure a max_disk_threshold of the postgresql-single-small service plan, to stop all processes on that node, when the disk usage exceeds 90% (the default value is 80%).

You need to adjust the service configuration like this:

postgresql-single-small:
[...]
default_placeholder_values:
[...]
max_disk_threshold: 90
migration_paths:
[...]