02
Published incidents
None. We have not declared an incident. This section lists them when we do, in the words described below, and it is empty rather than filled with an example.
03
How we tell you about an incident
An incident is any period where Laver is unusable, materially degraded, or has lost or exposed data. What we commit to:
- This page first. It is the single place we publish to, so there is never a question of where to look. We post here within an hour of confirming an incident.
- An update at least every two hours while an incident is open, even when the update is "no change yet". Silence during an outage is its own failure.
- Plain words about what is affected — which parts of Laver, and what you can and cannot do meanwhile. Not a cause, not a component name, and never a guess dressed as a diagnosis.
- Email to workspace owners for anything involving data loss or unauthorised access, rather than leaving it to be discovered here.
- A written summary within five working days of resolving anything that lasted longer than thirty minutes: what happened, what it affected, and what we changed. It stays on this page.
If you are affected and we have not posted, tell us at support@developyn.com — that is also the address to use if this page itself is unreachable.
04
How this page is measured
Every 30 seconds, Laver records that it is running and that its database answered. Those recordings are joined into continuous stretches of service; a gap between two stretches is counted as unavailable, and that is where every figure above comes from. Percentages are rounded down, so a real interruption can never be published as 100%.
We started measuring on 5 August 2026. Anything before that is shown as not measured, because we have no evidence about it either way.
What this cannot tell you
- A network failure between you and us. The recording is made on the same machine that serves Laver, so if that machine is healthy but unreachable from where you are, this page will show it as served. A prober elsewhere is the only fix, and it needs a second host.
- This page goes down with Laver. It is served by the same host it reports on. If a failure takes the whole host out, you will not be reading this — which is a real limitation of publishing it here, and the reason the email address above matters.
- Very short interruptions — under 90 seconds — may not appear. A single late recording is not treated as an outage, because publishing invented downtime is no better than hiding real downtime.
- Speed is not covered. This measures whether Laver answered, not how quickly. A slow but working Laver reads as served here.
How Laver is hosted, backed up and secured is set out in the Trust Centre.