1.3.0
🚀 PatchMon version 1.3.0
This major release brings a new and improved architecture on the server and agent.
✨ Major Features
GO based Agent
Agent is now a GO based Agent binary file that runs as a service systemctl status patchmon-agent
The Agent serves a websocket connection to PatchMon server.
Agent has been compiled in amd64, i386, arm and arm64
A new repository has been setup for the agent.
BullMQ + Redis
The PatchMon Server runs a BullMQ service which utilises Redis server for scheduling automated and queued jobs. Jobs include things like "Cleanup orphaned repos" where it will remove repositories that are now not associated with any hosts etc Bullboard has also been added so that we can have a dashboard to monitor the jobs from a server level.
WebSocket
PatchMon Agents now connect via Web Socket Secure (https) or Web Socket (ws) to listen for commands from PatchMon. The Agents themselves control the schedule of reporting information however this persistent and bi-directional connection architecture lays the foundation of PatchMon so that it can control and handle management etc.
Performance
Various performance related improvements have been made with the way that node.js uses prisma for the Postgresql ORM. There was a lot of connection leakage where instead of utilising established connections it would create a new connection to the Database. These were causing at times Database connections to rise above 100! Fixes also improved the way the front-end speaks to the /api/v1 endpoints. These remove and handle the 429 (rate limit errors) and other backend errors.
Security
Various security handling has been improved around cookie handling, cors handling etc so that /bullboard can be authenticated
Agent updates checking
New mechanism for checking for Go based agents. The PatchMon server will query the GitHub repo and allow you to download the agents directly whilst the agents themselves will query PatchMon. I have pinned the agent version with the server version and had the agents query the server for downloading the updates as opposed to downloading them from github. This is because I plan to use PGP for signing agents off in the future and improve security mechanisms for server/agent verification.
Upgrading
Bash scripts from 1.2.8 will use an intermediary script of 1.2.9 which will run the installation of the new agent service. Docker upgrade instructions video : https://www.youtube.com/watch?v=NZE2pi6WxWM
Coming soon: For bare-metal type the setup.sh update is being modified soon to handle the installation and setup or Redis 7 DB user and password as well as the nginx configuration amendments to handle upgrade on the websocket and add the /bullboard directive.
Many thanks to the community for their hard work and support. <3 iby___