About Cahaves
Service health.
Every service the house runs, and whether it is up right now. If something here disagrees with what you are seeing, trust what you are seeing and tell us.
All systems normal
- operational
Cloud
Deploys, hosting and the dashboard
- operational
Studios
Project sync and asset delivery
- operational
Tech
Support desk and scheduling
- operational
Accounts
Sign in, billing and SSO
- operational
Docs and help
Documentation and the help center
These states are hard-coded and do not yet come from a monitor. A status page that cannot go red is worse than none at all, so treat this as a layout until it is wired to a real source.
What up means here
Up means the service is doing the thing it exists to do, for the people using it, at this moment. It does not mean a process is running or that a health check answered. A deploy queue that never drains is not up because the machine holding it replied.
So the check is made from outside, on the paths people actually take: a request served, a build finished, a sign-in completed. If those work and something behind them does not, the row still says degraded, because the part that failed is the part somebody needed. A green dot means nothing until you know what turns it another colour.
Degraded is not down
The two words are not interchangeable, and which one is used decides what you should do next.
- Degraded
- The service answers, but not as it should. Slower than usual, one region affected, or one operation failing while the rest hold. Work still gets done, with friction, and a retry is often worth trying.
- Down
- The service is not doing its job for the people using it, and retrying will not change that. A degradation nobody can work around is treated as an outage whatever the graph says: the label follows what you can do, not what the monitor reports.
How an incident is communicated
This page first, because it is the one place people are told to check. An open incident gets an entry here while it is still running, and the entry says what is affected, what we know, what we are doing about it, and when the next update lands.
Updates come on the stated cadence whether or not there is news, so silence never has to be interpreted. When it closes we write what happened and what changed as a result. Anyone with an open engagement also hears from the person who owns their account. That is on top of this page, not in place of it.
If you see something this page does not
Trust what you are seeing. This page is a report, and a report can be wrong or late. If a service is failing for you while its row above reads operational, tell us: what you did, what happened, when, and the region you were in if you know it.
If what you found is a security issue rather than an outage, do not send it to support. The disclosure policy has the address and says what happens after the report arrives.