Monitoring¶
This section describes the monitoring to be set up on a Reemo platform
deployed with the reemo-infra role: what to check, which probes to
install, and which commands to run when an alert is triggered.
The checks depend on the installation type and the enabled options (Workstation, appliances, credential portal, TURN, Vault, backups, etc.).
Choose your installation type¶
A single infra_manager cluster hosts the database, API, and portal.
Separate api_manager, portal_manager, and, if needed, turn_manager clusters.
In both cases, relayws_manager (Relay WebSocket) and provisionN_manager
(Reemo Containers) clusters can complement the installation.
Machine profiles¶
Each profile corresponds to an Ansible inventory group and an independent Docker Swarm cluster.
Profile |
Hosted services |
|---|---|
|
All-in-one: database, API, portal, signal, traefik, and depending on options: Workstation, appliances, credential portal, TURN, Vault, backups. |
|
Database, API, proapi, prorelayapi, credentialapi, applianceapi, logapi, Vault, backups, traefik (mTLS required). |
|
Portal, administration portal, signal, Workstation, appliance portal, credential portal, traefik, and an haproxy proxy to the API. |
|
Dedicated TURN server and traefik. |
|
Relay WebSocket, traefik, nginx. |
|
Container provisioning machines, nginx. |
Conventions¶
The examples use
INSTANCE_NAME=reemo: Docker Swarm services are therefore namedreemo_api,reemo_portal, etc. Replace this prefix with your instance name.Ports are those of the role’s default variables. Adapt them if you have modified them in your inventory.
Probes are written for Nagios and transpose to any tool that uses the same return codes (Icinga, Centreon, Naemon, Zabbix via an agent, etc.).
See also¶
Monitoring an all-in-one installation — checks for an
infra_managerinstallation.Monitoring a Split Installation — checks by profile for a split installation.
Monitoring Checks — details of each check point, component by component.
Nagios Probes — probe scripts and Nagios / NRPE configuration.
Quick status check — status overview in minutes when an issue occurs.
Alert Troubleshooting — diagnostic commands for each alert.