Observability
Know what broke before anyone asks.
Live logs, incidents, traffic and findings the server spots on its own, with an Explain button wherever something went wrong.
Findings, found for you
The server notices a service down or short of replicas, failing tasks, the change made just before, certificates about to expire and a disk filling up. No model needed.
Logs that follow along
Live, with search, levels, highlights and a time range up to 30 days, per container or all together, exported as text, JSON or CSV.
Incidents, with a timeline
When an app goes down an incident opens, and closes when it recovers. Each app keeps its incidents and the history of its changes.
Told where you already are
Slack, Discord and Mattermost webhooks for deploys, starts and stops, an app going down, and approval requests.
In detail
What you get
Explain, where it happened
- A rollout that did not finish, a container killed for memory or a failed healthcheck, an error line, an incident: each has an Explain that opens Cosmo with the context filled in.
Watches
- Ask Cosmo to keep an eye on something: on a schedule from every 15 minutes to weekly, or when an app goes down.
- Watches only read, and can post to your notification channels.
And the rest
- An audit log of who did what, filterable and grouped by day.
- Container stats, inspect and logs.
- The dashboard: services, CPU, memory, requests, error rate and firewall drops at a glance.
With Cosmo
Every finding on the dashboard opens its own conversation: Cosmo reads the tasks, the logs and the last change, says why, and puts the fix on the same card for you to approve.
How Cosmo works- “Is anything down or restarting over and over?”
- “Summarise checkout’s errors in the last hour”
- “Watch shop and tell me if it goes down”
Your servers are waiting.
One command, a minute or so, and the panel is on your domain. Free, and yours to keep.
curl -fsSL https://coswarm.dev/install.sh | bash