This is a very basic docker container that healthcheck the nodes of a Mariadb Galera Cluster.
Contributors :
Summary :
Use the healthcheck to monitor a specific and retrieve metrics from Galera. To determine the health of your cluster, the script will monitor :
- The number of connected node in the cluster
- The state of the cluster
- The state of your node inside the cluster
- The sync status of your cluster
This metrics are fetch from a node, which mean that you can detect network partition, malfunction, lose of quorum or desync. Thus, to have a precise and global view of your cluster, it is recommanded to healthcheck every node of your cluster.
For now, the only mean of notification implemented is through discord webhook notification.
When the cluster is healthy :
When an error is detected :
Thus the script is made to be modulable with many other notification medium.
You can use the file from the repo, or directly this config :
version: '3.7'
services:
galera_health_check:
container_name: galera_health_check
image: ghcr.io/seaweedbraincy/galera_cluster_healthcheck:latest
user: 1000:1000
security_opt:
- no-new-privileges:true
read_only: true
environment:
DB_USER: <REPLACE_ME> # required
DB_PASSWORD: <REPLACE_ME> # required
DB_HOST: <REPLACE_ME> # required
DB_PORT: 3306 # optional. Default is 3306
CHECK_INTERVAL: 60 # optional. Default is 60. Healthcheck interval in seconds
ALERT_THROTTLE: 10800 # optional. Default is 10800s (3h). Amount of time in senconds between 2 consecutive alerts
NODE_NAME: <REPLACE_ME> # required. Used to identify the node in the alert message
CLUSTER_MIN_SIZE: <REPLACE_ME> # required. Minimum number of nodes in the cluster, used to determine if the cluster is in a degraded state
WEBHOOK_URL: <REPLACE_ME> # required. Discord webhook URL for notifications
ROLE_TO_MENTION: <REPLACE_ME> # option. Discord role ID to mention in the alert message. If not provided, no role will be mentioned
volumes:
- ./docker_logs:/app/logs
restart: alwaysYou can use several env variable to configure the script :
| Variable | Is required | Defintion |
|---|---|---|
| DB_USER | REQUIRED | Mariadb user used to connect to the database. The user should only be able to read mariadb GLOBAL STATUS variables |
| DB_PASSWORD | REQUIRED | Password of the user used to connect to the database. |
| DB_HOST | REQUIRED | Mariadb host |
| DB_PORT | optional | Port used by mariadb. Default value : 3306 |
| ALERT_THROTTLE | optional | Amount of time in senconds between 2 consecutive alerts. Default value : 10800s (3h). |
| CHECK_INTERVAL | optional | Healthcheck interval in seconds. Default value: 60 |
| NODE_NAME | REQUIRED | Used to identify the node in the alert message. |
| CLUSTER_MIN_SIZE | REQUIRED | Minimum number of nodes in the cluster. This is used to determine if the cluster is in a degraded state |
| WEBHOOK_URL | REQUIRED | Discord webhook URL used to send notifications |
| ROLE_TO_MENTION | optional | Discord role ID to mention in the alert message. If not provided, no role will be mentioned |
This healthchecker monitor the 5 following variables:
- wsrep_cluster_size
- wsrep_cluster_status
- wsrep_local_state_comment
- wsrep_connected
- wsrep_incoming_addresses
In case of error with any of them, all the variable checks are reported in the notification message. Those one in error, would have the 'ERROR' mention, with a quick explanation of the issue.
You can find further troublehsooting variable and more information in the .
Any time the healthchecker finds an error (while checking or if the check fails), you will recieve a notification.
If and only if all checks pass, you will recieve a notification of normal state. While you don't have any green notification following the error notification, that means that you still have to investigate an issue.
All errors and messages are logged into Docker logs, they are not throttles. You can use :
docker compose galera_cluster_healthcheck logsAll contributions are welcome ! Feel free to open a merge request
The code is under MIT Licence