System Status
Real-time health of FrontlineHQ services.
Core Infrastructure
AI Services
Integrations
What We Monitor
Database (Supabase)
Our PostgreSQL database on Supabase powers all business data, conversations, leads, contacts, and knowledge bases. We monitor connection latency, query performance, and replication health every 60 seconds. Row-level security ensures each business can only access its own data. If database performance degrades, chatbot responses, dashboard loads, and lead capture may slow down.
AI Chat (Anthropic)
FrontlineHQ uses Anthropic's Claude AI as the primary language model for chatbot conversations, review responses, and lead qualification. We check API availability and response latency continuously. If Claude is unavailable, our fallback chain automatically switches to backup models to keep your chatbot and voice agent running without interruption.
Cache (Redis)
Redis powers our rate limiting, session management, and response caching layer. This keeps the platform fast and protects against abuse. We monitor cache hit rates and connection health. If Redis is degraded, we fall back to in-memory rate limiting so your chatbot and voice agent remain operational.
Payments (Stripe)
Stripe handles all subscription billing, trial management, and payment processing. We verify webhook connectivity and API availability so that subscription changes, upgrades, and new signups work smoothly. Payment data never touches our servers — it goes directly to Stripe.
Voice Agent (Retell)
Retell AI powers our voice agent — the AI receptionist that answers phone calls for your business. We monitor call routing, agent availability, and transcription services. If Retell experiences degradation, incoming calls may route to your fallback number or voicemail until service is restored.
Search Re-ranking (Cohere)
Cohere's reranking model improves the accuracy of chatbot responses by sorting search results by relevance before they reach the AI. We monitor reranking latency and quality scores. If Cohere is unavailable, chatbot responses still work using our base vector search — responses may be slightly less precise but remain functional.
How Our Monitoring Works
FrontlineHQ runs automated health checks every 60 seconds against all critical services. Each check measures connection time, response latency, and functional correctness. Results are aggregated into three status levels: Operational (all checks pass), Degraded (some latency or partial failure), and Down (service unreachable).
When a service enters a degraded or down state, our engineering team is alerted immediately via Sentry and Slack. Most issues are resolved within minutes through automated failover. For extended outages, we post updates on this page and notify affected customers by email.
Our architecture is designed for resilience. The AI chatbot has a three-model fallback chain, the cache layer has in-memory failover, and the voice agent can route calls to your backup number if needed. This means your customers experience minimal disruption even when individual services have temporary issues.
Need Help?
If you are experiencing issues that are not reflected on this status page, please contact our support team. We are available by email at support@frontlinehq.ai and by phone at (855) 677-7034.
For product updates and new features, visit our changelog. To learn more about what FrontlineHQ can do for your business, explore our AI solutions or read the about page.