New The SREYUN website is live: five in-house products and reference pricing Learn more →

openaiops Agent OTA: the fleet keeps up after a server upgrade

Product releases By SREYUN 4 views

Once the fleet grows, upgrading agents is often harder than upgrading the server: replacing binaries over SSH host by host always misses a few or fails on some. openaiops Agent OTA makes this a built-in capability of the platform.

Automatic and controlled

  • Automatic OTA (on by default): online agents that are behind are queued automatically when they report in; you can also push in batches or to selected hosts from the console.
  • Safe switchover: the binary for the right OS and architecture is fetched from the server and checked with SHA-256; a backup is taken before switching and failures roll back automatically.
  • Controlled rollout: maintenance windows, exemptions by host or category, and a circuit breaker after repeated failures with the same cause; you can see why any host was not upgraded.

Visible end to end

Upgrade tasks move from running → pending_verify → success / failed in plain sight; Windows, Linux and macOS each have their own upgrade helper and logs, and the console can pull evidence from the host.

Deployment note

OTA and the remote terminal share the agent's reverse connection. If there is an Nginx reverse proxy in front of the server, turn off buffering in both directions and allow WebSocket, otherwise you get "metrics fine, upgrades failing"; the project documentation has a reference configuration.

For everything openaiops can do, visit the product page.

Related articles