Product4 min read
Introducing Watches: scheduled agents for the quiet parts of operations
Watches send Keel's agents through your stack on a schedule, looking for slow-burning risks and clearing the recurring chores that never make it onto a sprint.
By Ines Carvalho
Most operational risk does not arrive as an outage. It accumulates: a certificate that expires next month, a replica that drifts a little further behind every week, a staging cluster nobody has touched since spring. None of it is urgent enough to page anyone, so it waits until it is.
Today we are introducing Watches, scheduled agents that walk through your environment looking for exactly this kind of slow-burning problem.
How a Watch works
- Each Watch has a focus, such as cloud spend, database health or monitoring coverage, and a schedule that suits it.
- On every run it checks the parts of the System Map in its scope and compares what it finds with what it saw last time.
- Anything worth acting on becomes an Incident or an Optimization in your queue, with the evidence attached and, where it can, a drafted fix.
The first seven
We are launching with seven Watches that cover the areas our early customers asked about most: spend, monitoring coverage, database health, latency, capacity, expiring credentials and error sweeps. Each one can be scoped to specific accounts, services or repositories, and each one learns from the decisions your team makes about its findings.
If your team tells a Watch that a finding is intentional, it remembers, and it will not raise the same thing again unless something changes.
Watches are available today in every Keel workspace. Turn on your first one from the Watches page, or ask Keel in your incident channel to set one up for you.
