Ansible role: Beszel monitoring agent (systemd, Homebrew or Docker)
- Jinja 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
- launchd method replaces brew: the formula builds from source when there is no bottle for the macOS version, and its tap needs trusting. The release binary is extracted with tar (unarchive can't use bsdtar) and the log is owned by the agent's user, or launchd fails with EX_CONFIG. - docker method uses the docker CLI: Unraid's Python lacks requests. - Unpack again whenever the binary is missing. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
| defaults | ||
| handlers | ||
| meta | ||
| tasks | ||
| templates | ||
| README.md | ||
ansible-role-beszel-agent
Installs the Beszel agent. It listens on
beszel_agent_port (45876) and accepts only the hub whose public key is
beszel_agent_key; the hub connects to it over SSH.
beszel_agent_method |
Where | How |
|---|---|---|
systemd (Linux default) |
Debian/Ubuntu | Pinned release binary (checksum-verified) in /opt/beszel-agent, systemd unit modelled on upstream's, beszel user in the docker group when present |
launchd (macOS default) |
Macs | Pinned release binary (checksum-verified) in /opt/beszel-agent, LaunchDaemon dev.beszel.agent running as the login user. Not the Homebrew formula: it builds from source when there is no bottle for the macOS version. |
docker |
Unraid | henrygd/beszel-agent container (docker CLI, no Python SDK needed), host network, Docker socket read-only |
Extra agent settings go in beszel_agent_extra_env.