How we run in production.
Orbyt Intelligence does not publish a contractual SLA today. Here is the honest version: the infrastructure we run on, how we handle incidents, and what we'll commit to in writing when we're ready to back it with credits and termination rights. We'll say so on this page the day that changes.
Orbyt is a solo-founder product. We do not have a written SLA with service credits or termination rights. We operate on infrastructure (Vercel, Supabase, Stripe) that is itself SOC 2 Type 2 certified, but our own SOC 2 audit is not yet underway. We treat reliability as a first-class engineering concern: but we will not promise uptime numbers we cannot contractually defend. When that changes, this page changes with it.
Infrastructure we run on.
Orbyt's reliability is largely inherited from the platforms underneath. The providers we use publish their own SLAs and SOC 2 reports: and we lean on them rather than reinventing the operational stack.
Orbyt's own code is the layer on top: request routing, authentication, the data engines, the MCP server. We deploy from a single git main branch to Vercel. Every PR is reviewed and tested in CI before merge. Rollback is a one-command revert.
What we monitor.
We watch a few things continuously. None of these surface as a public status page yet , that's on the roadmap. Until it lands, customers who experience an issue can email support@orbytjobs.ai with the affected request_id from the response envelope and we can trace it in Sentry within minutes.
- Error rate across all Intelligence endpoints (Sentry alerts above 1% 5xx for 5+ minutes)
- Response time: Vercel surfaces p50, p95, p99 latency per route; alerts on p95 above 2× baseline
- Auth failures: sustained 401 spikes could indicate a key leak or middleware regression
- Database health: Supabase dashboard for query latency, connection pool, replication lag
- Stripe webhook delivery: failed webhook events are caught and replayed; Sentry alerts on sustained delivery failures
Incident handling.
When something breaks, the goal is: detect, communicate, fix, learn. We don't have a 24/7 on-call rotation: Orbyt is a solo product today. But the response process is the same shape larger teams use, scaled down:
- Detect. Sentry alerts to a Slack channel. Vercel deployment health and Supabase status dashboards are checked daily.
- Communicate.If customer-visible, affected Pro+ customers get an email with what we know, what's being done, and an ETA. We do not have a public status page yet, so direct email is the channel.
- Fix. For most incidents the fix is a code rollback (revert the offending PR and redeploy: ~3 minutes on Vercel) or a Vercel/Supabase status-page wait when the issue is upstream.
- Learn. Material incidents get a short postmortem captured internally. Patterns that show up more than once become proactive monitoring rules.
If you hit a problem, email support@orbytjobs.ai with the response's request_id (every Orbyt response includes one in the envelope). We can trace it end-to-end in seconds.
On the roadmap.
Things we plan to ship, in roughly the order we'll ship them. No dates committed , Orbyt won't put a date on this page until the work is actually scheduled.
- Public status page with real-time uptime, error rates, and historical incident archive. URL and exact format to be announced when it ships.
- Formal SLA for Enterprise customers with measurable uptime targets and service-credit policy. The number on the SLA will be one we can actually defend, not industry-standard marketing copy.
- SOC 2 Type 1 audit engagement.
- Customer-facing log explorer in the dashboard: every request, its
request_id, the resolved tier and scopes, and the response status. (Backend infrastructure is built; the frontend explorer is what's pending.) - Multi-region database failover. Supabase supports it; Orbyt's data volume hasn't needed it yet, but it's the next step for Enterprise readiness.
When any of these ship, this page changes. Customers can subscribe to the changelog to be notified.
See also.
Last updated May 2026. This page describes Orbyt's current operational practices. It is not a contractual commitment. When Orbyt publishes a formal SLA, this page will be updated and customers on Pro and above will be notified by email 30 days in advance.