Incidents and status
Where a Velor incident shows up, what we do and do not send you, and what to prepare on your side.
Stated as it is today, including what we do not do yet.
Where to look
- `/status` is the public status page, read live from our own checks on every load, not cached. It shows current state, month-to-date availability measured by our probes, and incident history. If it cannot read a source, it says it could not check; it does not print an all-clear.
- The dashboard shows a banner when we post a known issue. One issue is active at a time, it is at most 280 characters, it has a severity of info, degraded or outage, and it expires on its own within seven days.
- `/dashboard/health` (Health) shows the health measurements for your workspace, and links to the status page.
Two controls we may use, and what customers see
- AI off. We can switch AI answers off for the whole platform. Chat, voice, SMS and email then answer with a short holding message and conversations are handed to a person where you have a team. The dashboard, API and knowledge base keep working.
/statussays so in fixed wording. - Read-only maintenance. You can read the dashboard but changes are refused. Conversations, webhooks and the API are not affected.
If a person on your team is not online when AI is off, the customer receives the holding message and waits. This is why the going-live checklist asks for someone to be online.
What we do not do
- Alert emails are not sent. You can leave an address at
/status/subscribeand receive a confirmation email, but no incident notification is sent to subscribers today. The page says the same. Do not rely on it. Check/statusor your dashboard, and add our status page to your own monitoring. - There is no on-call phone line and no incident SLA in the product. Commitments, if you have any, are in your signed agreement, not on this page.
- We do not email every customer on an incident. If an incident affects you materially we contact the owner address on your account; that is manual.
Reporting a problem to us
Email caden@velorplatforms.com (the support address; there is no ticket queue), or security@velorplatforms.com for a security issue. Include the time, the channel, a conversation id from /dashboard/conversations if there is one, and what you expected.
What to prepare on your side
- Know your fallback. Decide now what your widget visitors should see if Velor is down, and whether your phone and SMS numbers have a Twilio-side fallback. Your numbers are on your own Twilio account, so the carrier-side fallback is yours to configure.
- Make webhook handlers idempotent. Deliveries are retried, so the same event can arrive more than once. Deduplicate on the payload
id, which stays the same across retries, and acknowledge duplicates. - Monitor us from your side. Poll
/api/healthor/statusfrom your own monitoring, and alert your team on your own channels. - Tell your customers in your own words. We provide no customer-facing incident template for your brand.
In the product: /dashboard/health (sign-in required).