Autonomous Agent
604 memories | avg score 26% | 9 agents | never deleted
The overall error rate in Cloud Run Metrics is high (6.9%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
REGLA DE CLASIFICACIÓN HIGH_MODERATION: Usuario con block_rate >10% Y/O report_ratio >5% (mínimo 10 conversaciones para calcular ratios) = clasificación high_moderation. Estos usuarios reciben escrutinio adicional en todos sus posts y mensajes. Usuarios SUPERBLOCK: NUNCA se mantiene contacto, se ignoran completamente todos los mensajes.
REGLA DE SPAM EN CHAT: >5 mensajes idénticos en 24h O >10 destinatarios únicos en 1 hora = Block Chat + evaluar Block User. Mensajes muy cortos ("Precio?", "Disponible?") enviados en masa = spam. Mensajes con promesas de dinero rápido, solicitudes de datos bancarios, o links sospechosos = Reject Chat inmediato.
REGLA DE DUPLICADOS: Cuando un usuario publica anuncios duplicados: Ghost el más antiguo, mantener el más reciente. Si solo publica duplicados = spammer → Block. Usuarios nuevos con duplicados válidos: aceptar 1, rechazar resto por DUPLICATE (sin bloquear al usuario).
REGLA DE VENDEDORES PROFESIONALES: Vendedores profesionales haciéndose pasar por particulares = Ghost (ocultar sin que el usuario lo sepa, NO reject directo). Indicadores: fotos de catálogo + texto genérico + múltiples anuncios similares. Escalar a revisión manual para confirmar antes de tomar acción definitiva.
REGLA DE CONTACTO EN ANUNCIOS: Información de contacto (teléfono, email, URL) en título, descripción o imagen del anuncio = needs_editing (NO reject). Dar al usuario la oportunidad de corregir. Solo el campo habilitado es válido para contacto. Excepción: datos sensibles como DNI/pasaporte en imagen = Reject > Prohibited directamente.
REGLA DE DETECCIÓN DE SCAM: La combinación de pago adelantado + urgencia + precio demasiado bueno = REJECT por scam_detected con alta confianza. Productos en "otra ciudad que requieren depósito para transporte" = scam. "Gana dinero rápido desde casa sin experiencia" = scam. Solicitud de PIN/código SMS = intento de robo de cuenta → Reject + marcar usuario sospechoso.
REGLA CRÍTICA: SIEMPRE escalar a humano: solicitudes de policía/autoridades/tribunales, reportes de robo de identidad confirmado, amenazas de violencia, usuarios que han sido estafados (necesitan seguimiento real), solicitudes GDPR de eliminación de datos, problemas con pagos/facturación, temas de marca registrada.
REGLA CRÍTICA: NUNCA aprobar anuncios de armas, drogas, órganos humanos, contenido con menores, productos robados, servicios sexuales o cualquier contenido que ponga en riesgo la seguridad física de las personas. Acción: Reject > Prohibited + Block User si es grave.
Platform Activity query failed; this needs investigation as it blocks monitoring of key business metrics.
SÍNTESIS: MODERACIÓN - ESTÁNDAR DE LA INDUSTRIA: Todos usan combinación de: revisión automática + manual. Community flagging/reporting. Sanciones progresivas (warning → temp block → permanent ban). Ninguna plataforma ofrece atención telefónica (todo escrito). Diferencias clave: Craigslist: mínima moderación, community-driven flagging anónimo, $4 multa por post prohibido. OfferUp: más estricto (firearms = NO exceptions, COVID vaccine posts = immediate account removal). Facebook: vinculado a identidad real, moderación de IA avanzada. Wallapop: políticas detalladas por categoría, Wallapop Pro para profesionales. Milanuncios: Express con verificación KYC, conectado con verticales (Coches.net, Fotocasa). OLX: políticas muy detalladas por producto (joyería con pureza, comida con origen de carne), portugués/español obligatorio. APLICACIÓN A SELIO: Nuestras políticas están alineadas con el estándar de la industria. Áreas donde podemos aprender: retención de pago (como Wallapop/Milanuncios Express), DAC7 fiscal compliance, departamentos de policía como punto de encuentro (Craigslist).
SÍNTESIS: REGLAS DE SEGURIDAD - ESTÁNDAR DE LA INDUSTRIA: TODOS los competidores recomiendan: Transacciones en persona en lugar público. No enviar dinero por adelantado. No compartir datos personales/financieros. No salir de la plataforma (WhatsApp/Telegram). Reportar comportamiento sospechoso. MEJORES PRÁCTICAS observadas: Craigslist: sugiere departamentos de policía como punto de encuentro. OfferUp: monitoreo 24/7 con equipo Trust & Safety + detección automatizada. Wallapop: Wallapop Envíos con retención de pago hasta confirmación. Milanuncios: Express con dinero retenido 24h, verificación de identidad obligatoria para vendedores, DAC7 fiscal reporting. Facebook: vinculación con perfil real de Facebook para verificación. SEÑALES DE ESTAFA comunes a todos: Precio demasiado bajo, urgencia, solicitud de códigos SMS, links externos, envíos a otra ciudad con depósito, ofertas de empleo irreales.
SÍNTESIS: ARTÍCULOS PROHIBIDOS - ESTÁNDAR DE LA INDUSTRIA (todos los competidores coinciden): UNIVERSAL (100% de plataformas prohíben): Armas de fuego y munición. Drogas ilegales y sustancias controladas. Medicamentos con receta. Contenido de explotación infantil. Servicios sexuales y trata. Artículos robados. Documentos oficiales falsificados. Productos falsificados/réplicas. Explosivos y fuegos artificiales. Partes del cuerpo humano y órganos. Animales vivos (la mayoría). Contenido de odio/discriminación. Esquemas piramidales/MLM fraudulentos. MAYORÍA (>80% prohíben): Tabaco y cigarrillos electrónicos. Alcohol. Artículos con contenido sexual explícito. Criptomonedas y servicios financieros. Dispositivos de vigilancia/interceptación. Gift cards y cupones digitales. Cuentas digitales (Netflix, Spotify). Contenido descargable (PDFs, ebooks). Mystery boxes sin descripción. VARIABLE (depende de plataforma): Productos de salud/suplementos (OfferUp prohíbe todos, OLX/Wallapop más permisivos). Mascotas con documentación (Milanuncios permite a profesionales, Craigslist permite rehoming). Piezas de vehículos (Wallapop prohíbe excepto centros autorizados, otros permiten). Artículos de colección militar/nazi (OLX permite sellos/monedas del III Reich, OfferUp prohíbe todo).
ESTÁNDARES DE MODERACIÓN DE COMPETIDORES: La industria converge en: verificación de identidad progresiva (todos), moderación automática + manual (todos), prohibición de armas/drogas/adultos (universal), protección de comprador (Wallapop, Vinted, eBay), community flagging (todos), detección de spam automática (todos). Selio se diferencia por: focus en mercados hispanohablantes/lusófonos, sistema de trust basado en eigenvector, moderación LLM con Gemini, two-rows-per-conversation para privacidad, Manticore Search para full-text + percolate queries.
ARQUITECTURA DEL SISTEMA DE CONVERSACIONES: Dos filas por conversación en Spanner (una por participante). conversation_id = MD5(sorted user IDs). Cada fila tiene estado independiente: folder, unread_count, blocked_at, reported_at, rated_at, would_deal_again. Bloqueo = inline (no tabla separada). Rating = inline (no tabla separada). Email relay: {nombre}.{código16hex}@seliovibe.com. Firestore sync signal: documento sync/{user_id} con timestamps (conversations_at, messages_at, notifications_at, saved_at). Flutter detecta cambio → fetch desde API. Spanner = fuente de verdad única.
CLASIFICACIÓN Y TRUST SCORING DE USUARIOS: Clasificación automática basada en: edad de cuenta (NEW <7 días), ratios de bloqueo/reporte (HIGH_MODERATION si block >10% o report >5%, con mín 10 convs), actividad (ACTIVE/CHURNED), verificación profesional (PROFESSIONAL), historial excelente (TRUSTED: >90 días, >5 posts, >10 ratings positivos, >50% deal rate). Trust levels: NEW (0-20), LOW (0-30), MEDIUM (30-60), HIGH (60-85), VERIFIED (85-100). Cada nivel determina el grado de moderación aplicado a los posts del usuario.
RECONOCIMIENTO DE PATRONES DE ESTAFA: Patrón clásico: precio demasiado bajo + urgencia + pago adelantado = scam. Patrón phishing: solicitud de PIN/código SMS + "verificar anuncio" = robo de cuenta. Patrón 419: príncipe/herencia/lotería + transferencia internacional = scam nigeriano. Patrón reubicación: "producto en otra ciudad" + depósito para transporte = estafa. Patrón empleo: "gana 500€/día desde casa" + sin requisitos = fraude. Patrón romance: declaraciones emocionales rápidas + solicitud de dinero = romance scam. Patrón producto alterado: "ese ya lo vendí, pero tengo otro" = bait-and-switch.
REGLAS DE DUPLICADOS Y SPAM: Duplicados del mismo usuario: aceptar 1 (el mejor), rechazar resto por DUPLICATE. Si solo publica duplicados = spammer → Block. Auto-renovación abusiva: máximo 1 vez al mes. Bait-and-switch (publicar algo y luego cambiar) = Ghost. Competencia directa: no permitir anuncios que promocionen competidores. Normas: no dropshipping, no anuncios cebo, no servicios ilegales disfrazados de legítimos.
IDENTIFICACIÓN DE VENDEDORES PROFESIONALES: Señales: >5 anuncios activos del mismo tipo de producto, fotos de catálogo en todos los anuncios, texto genérico/plantilla repetida, patrón de publicación comercial continuo. MLM: Herbalife, Tupperware, Oriflame, Avon = rechazar si es recurrente/comercial (permitir si es puntual de particular). Cuentas Netflix/HBO/Spotify = Reject > General. Acción: escalar a revisión manual para confirmar. Si profesional no registrado → no bloquear directamente, evaluar caso por caso.
PATRONES DE EVASIÓN DE BAN: Los usuarios bloqueados intentan volver con: nuevo email (detectado si email linked to blocked), nuevo teléfono (phone linked check), mismo dispositivo (device ID match), misma cara (face vector matching). Señales adicionales: disposable email domain, VPN/proxy IP, IP country mismatch con phone. Sistema de detección: verifica en registro, verificación de teléfono, verificación facial, y batch diario. La tabla blacklist almacena: tipo (email/phone/device/ip), valor, nivel de severidad, motivo, expiración opcional.
GUÍA DE SEGURIDAD PARA COMUNICACIONES: Para usuarios: reunirse en lugar público, verificar producto antes de pagar, no compartir datos bancarios/PIN/códigos SMS, no hacer pagos adelantados, usar chat de la plataforma (no WhatsApp), cambiar contraseñas regularmente, reportar comportamiento sospechoso. Para staff: Selio NO interviene en transacciones (solo tablón de anuncios). Selio NUNCA solicita datos por mensaje directo. Selio colabora con policía si hay denuncia. Los usuarios pueden ocultar teléfono en configuración de perfil.
GUÍA DE PIPELINE DE MODERACIÓN: Post creado → status PENDING → Cloud Task → Tier 1 (reglas rápidas, 0ms LLM): contacto en descripción, caps en título, descripción corta, verificación de teléfono. Si bloquea → decision directa. Si no → Tier 2 (LLM Gemini Flash): análisis de texto (8 scores 0-1) + imágenes. Thresholds: >0.7 reject, 0.3-0.7 review, <0.3 approve. Si review → Tier 3: cola de staff priorizada. Post-decisión: needs_editing → checks horarios 7 días → auto-reject. Blocked → checks horarios 30 días. Quality: 1% audit sample, 7-day outcome tracking.
REFERENCIA RÁPIDA DE CÓDIGOS DE MODERACIÓN: DECISIONS: approve (publicar), needs_editing (corregir), blocked (esperando verificación/staff), reject (definitivo), ghost (oculto pero usuario cree publicado), reject_ban (rechazo + suspensión). EDIT ISSUES: 22 códigos en 5 categorías (precio, fotos, título, descripción, categoría/ubicación). REJECTION: 18 códigos en 6 categorías (política, fraude, spam, off-platform, cuenta, IP). BLOCK REASONS: 8 códigos (phone/age/identity required, staff/senior/legal review, cooldown, daily_limit). USER CONSEQUENCES: none, warning, trust_penalty, rate_limit, shadow_ban, suspend_temp, suspend_perm.
PLAYBOOK DE RESPUESTAS DE SOPORTE: Tono: cálido, profesional, empático. Comenzar con "¡Hola!" y agradecer. Terminar con "Un saludo, El equipo de Selio". Escenarios clave: (1) Anuncio rechazado: explicar motivo específico, dar pasos para corregir. (2) Cuenta bloqueada: informar motivo, preguntar si quieren reactivar. (3) Fraude por WhatsApp: pedir número + capturas, recomendar ocultar teléfono, sugerir denuncia policial. (4) No ven su anuncio: puede tardar 24h, pedir que espere y reintente. (5) Estafado: derivar a humano siempre. (6) Policía/legal: derivar a humano siempre. (7) GDPR eliminación datos: derivar a humano siempre. (8) SUPERBLOCK: NO responder nunca.
ÁRBOL DE DECISIÓN DE MODERACIÓN DE CHAT: 1-¿Usuario SUPERBLOCK? → Ignorar. 2-¿Primera conversación + usuario nuevo? → Vigilar de cerca. 3-¿Contiene datos sensibles (PIN/DNI/contraseñas)? → Reject Chat + marcar usuario. 4-¿Fraude financiero (pagos urgentes/depósitos/transferencias)? → Block Chat. 5-¿Ganancias rápidas/métodos secretos? → Reject Chat. 6-¿Producto diferente al anunciado? → Reject Chat. 7-¿Empleo irreal con altos ingresos? → Reject Chat. 8-¿Documentos falsos/piratería/órganos? → Reject Chat + Block User. 9-¿Manipulación emocional (loterías/herencias)? → Reject Chat. 10-¿Insultos/amenazas/sexual? → Reject Chat (reincide = Block User). 11-¿Links redes sociales/QR/VoIP? → Reject Chat. 12-¿Spam masivo (idénticos/alta velocidad)? → Block Chat + Block User.
GUÍA DE EVALUACIÓN DE RIESGO DE USUARIO: Signals de alto riesgo: phone/device/email linked to blocked user, face match with blocked user, disposable email, VPN/proxy IP, IP country ≠ phone country. Velocidad sospechosa: >10 recipients/1h, >5 identical messages/24h, very short first messages en masa. Comportamiento bot: high night_activity_ratio, low profile_completeness, text_pasted (no typed). Historial negativo: block_rate >10%, report_ratio >5%, high post_removal_rate. Señales positivas: phone verified, face verified, corporate email, high deal_rate, good response_rate, camera photos (not gallery).
ESTÁNDARES DE CALIDAD DE POSTS: Fotos deben ser reales del producto (no catálogo, no otros sitios, no stock). Sin marcas de agua de competencia, sin QR, sin datos personales en imagen. Título descriptivo sin ALL CAPS, sin keyword spam, sin contacto. Descripción >20 chars sin info de contacto ni links externos. Precio realista (sin ficticios como 1€), sin decimales. Categoría correcta. Empleo: debe incluir salario + tipo contrato + descripción del puesto. Flujo de corrección: Edit > Photo/Information primero, si no corrige en 7 días = auto-reject.
POLÍTICA COMPLETA DE ARTÍCULOS PROHIBIDOS: Las categorías principales son: (1) Armas, munición y explosivos de todo tipo. (2) Drogas ilegales y medicamentos con receta (incluye esteroides, opioides, ansiolíticos, viagra). (3) Contenido adulto/sexual/obsceno. (4) Falsificaciones y propiedad intelectual (IPTV pirata, réplicas explícitas). (5) Animales sin documentación o protegidos. (6) Órganos y fluidos humanos. (7) Productos financieros no regulados y préstamos (SIEMPRE fraude). (8) Dispositivos de interferencia/radar. (9) Móviles bloqueados por IMEI. (10) Equipamiento estatal. (11) Productos químicos peligrosos. (12) Contenido de menores. (13) MLM recurrente. (14) Cuentas digitales (excepto Fortnite/dominios). Acción por defecto: Reject > Prohibited. Casos graves: + Block User.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services.
The overall error rate in Cloud Run Metrics remains high (7.6%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Cloud Run Health check failures for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner persist, correlating with high error rates in Cloud Run Metrics for these services.
The overall error rate in Cloud Run Metrics is high (8.2%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
Platform Activity, Business Metrics, Crawler Health, and Provider Budgets checks are failing due to database query errors, suggesting a potential database schema or migration issue.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates as observed previously.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates as observed previously.
Correlation between Cloud Run Health check failures and high error rates in specific services (selio-vibe-conversation-analyser, selio-vibe-crawler-agents, selio-vibe-scale-miner).
High error rates in selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner correlate with the Cloud Run Health check failures for those services.
Multiple CRITICAL alerts across Cloud Run Health, Cloud Run Metrics, and E2E Tests suggest a significant underlying issue.
The Sentry API error (403) indicates a problem with the Sentry token, potentially impacting error monitoring.
This is the first time we've seen Cloud Run Health, Cloud Run Metrics, and E2E Tests all reporting CRITICAL errors simultaneously.
Multiple CRITICAL alerts indicate a widespread issue affecting core services.
Overall Platform Stability: Declining (persistent critical errors)
Platform Activity: Stable (but at zero due to database issues)
Error Monitoring: Declining (Sentry API non-functional)
Service Health: Declining (significant drop in healthy services)
Correlation between Cloud Run Health check failures and high error rates in specific services (selio-vibe-conversation-analyser, selio-vibe-crawler-agents, selio-vibe-scale-miner).
E2E tests continue to fail, indicating ongoing issues with core functionality.
Platform Activity query failure persists, hindering the ability to monitor key business metrics and detect trends.
High error rates in Cloud Run Metrics correlate with the services experiencing health check failures, confirming the link between service health and error rates.
The Sentry API error (403) continues, preventing proper error monitoring and potentially masking underlying problems.
Cloud Run Health check failures persist, with selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner consistently failing, indicating a systemic issue.
E2E tests are consistently failing.
The Platform Activity query is failing, blocking monitoring of key business metrics.
The Sentry API is returning a 403 error due to an invalid or insufficient token, preventing proper error monitoring.
Persistent Cloud Run health check failures and high error rates are affecting selio-vibe-web, selio-vibe-conversation-analyser, selio-vibe-scale-miner, and selio-vibe-crawler-agents.
E2E tests continue to fail, indicating a potential regression in core functionality.
Platform Activity query failure persists, blocking monitoring of key business metrics. This is the second consecutive occurrence.
High error rates in selio-vibe-crawler and selio-vibe-crawler-agents correlate with Cloud Run Health check failures, suggesting a dependency issue or resource contention.
The Sentry API error (403) remains unresolved, preventing proper error monitoring. This is the second consecutive occurrence.
Cloud Run Health check failures persist for selio-vibe-web, selio-vibe-conversation-analyser, selio-vibe-scale-miner, and selio-vibe-crawler-agents, indicating a continued underlying issue.
E2E tests continue to fail, indicating a potential problem with core functionality.
The overall error rate in Cloud Run Metrics remains high (4.2%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a persistent issue.
Platform Activity query continues to fail, preventing monitoring of key business metrics. This is the second consecutive failure.
The Sentry API is still returning a 403 error, suggesting the token issue has not been resolved. This impacts our ability to effectively monitor errors.
Cloud Run Health check failures persist for multiple services (selio-vibe-conversation-analyser, selio-vibe-crawler-agents, selio-vibe-scale-miner), correlating with high error rates in Cloud Run Metrics, indicating an ongoing underlying issue.
The overall error rate in Cloud Run Metrics is high (2.8%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
Spanner CPU usage is at 60%, which is at the warning threshold. This should be monitored closely to ensure it doesn't reach the critical threshold.
The Platform Activity query continues to fail, preventing monitoring of key business metrics; this needs urgent investigation.
The Sentry API is still returning a 403 error, indicating a problem with the Sentry token that needs to be addressed to restore proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are still failing, indicating a persistent problem with the core functionality of the platform.
Spanner CPU usage is at 60%, which is at the warning threshold. While not critical, it warrants monitoring in subsequent runs to ensure it doesn't escalate.
The Platform Activity query continues to fail, preventing monitoring of key business metrics; this needs to be resolved to understand platform trends.
The Sentry API is still returning a 403 error, suggesting the token is still invalid or lacks the necessary permissions; this is impacting error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a continuing underlying issue.
E2E tests are failing, indicating potential problems with core functionality.
The overall error rate in Cloud Run Metrics is high (7.3%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query continues to fail, preventing monitoring of key business metrics and requiring investigation.
The Sentry API is still returning a 403 error, indicating a problem with the token that needs to be resolved to ensure proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are failing again after failing yesterday, indicating a regression or ongoing problem with the test environment or core services.
The overall error rate in Cloud Run Metrics remains high (8.7%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
Platform Activity query continues to fail, preventing monitoring of key business metrics; this is the second consecutive failure and requires investigation.
The Sentry API is still returning a 403 error, meaning error monitoring is still impaired; this needs to be addressed urgently.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics, indicating a persistent underlying issue.
E2E tests are failing consistently, indicating a problem with the application's core functionality.
The overall error rate in Cloud Run Metrics is high (10.4%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query continues to fail, preventing monitoring of key business metrics and requiring investigation.
The Sentry API is still returning a 403 error, indicating a problem with the token that needs to be addressed to restore error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are still failing, indicating ongoing issues with the platform's core functionality.
The overall error rate in Cloud Run Metrics is high (9.7%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics.
The Sentry API error (403) continues, indicating a problem with the Sentry token that needs to be addressed to restore proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are still failing, indicating ongoing issues with core functionality.
The overall error rate in Cloud Run Metrics remains high (7.4%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query continues to fail, preventing monitoring of key business metrics; this needs to be resolved to track post creation and moderation queue status.
The Sentry API error (403) persists, suggesting the token is still invalid or lacks the necessary permissions; this is blocking proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a continued underlying issue.
E2E tests are failing, indicating potential functional issues within the platform.
The overall error rate in Cloud Run Metrics remains high (7.1%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a persistent trend.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics. This is the second consecutive failure.
The Sentry API error (403) continues, preventing proper error monitoring. This is a recurring issue that needs resolution.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a systemic issue.
The overall error rate in Cloud Run Metrics remains high (7.0%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics. This is the second consecutive failure.
The Sentry API is still returning a 403 error, indicating a problem with the token. This prevents proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services. This is a recurring issue.
E2E tests are failing, indicating a problem with the application's functionality.
The overall error rate in Cloud Run Metrics remains high (7.2%), driven by errors in selio-vibe-crawler and other services experiencing health check failures, indicating the underlying issue has not been resolved.
Platform Activity query failed again, preventing monitoring of key business metrics; this issue persists from the previous run.
The Sentry API is returning a 403 error, indicating a problem with the token, which is preventing proper error monitoring.
Cloud Run Health checks are failing for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics, continuing a trend from previous runs.
Spanner CPU usage is at 59%, which is within normal limits, but should be monitored as it's approaching the warning threshold of 60%.
E2E tests are still failing, indicating ongoing functional issues within the platform.
The Platform Activity query continues to fail, preventing monitoring of key business metrics. This is the second consecutive report of this issue.
The Sentry API is still returning a 403 error, suggesting the token is still invalid and error monitoring is compromised. This is the second consecutive report of this issue.
Cloud Run Health checks continue to fail for selio-vibe-crawler, selio-vibe-conversation-analyser, selio-vibe-scale-miner, and selio-vibe-crawler-agents, indicating a persistent underlying issue. This aligns with the high error rates observed in Cloud Run Metrics for these services.
E2E tests are failing, indicating potential issues with core functionality.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics.
The overall error rate in Cloud Run Metrics remains high (7.7%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Sentry API is still returning a 403 error, indicating a problem with the token; this needs to be resolved to ensure proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services; this is a recurring issue.
Spanner CPU usage is at 63%, which is above the warning threshold of 60%, but still below the critical threshold of 80%.
Platform Activity query continues to fail, preventing monitoring of key business metrics; this is the second consecutive failure.
The Sentry API is still returning a 403 error, indicating a persistent problem with the Sentry token, which is impacting error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating an ongoing issue.
E2E tests are still failing, indicating ongoing functional issues within the platform.
Spanner CPU usage is at 61%, exceeding the warning threshold of 60%, but storage remains low; this could indicate a performance bottleneck.
Platform Activity query failed again, preventing monitoring of key business metrics; this is the second consecutive failure and needs investigation.
The Sentry API is still returning a 403 error, suggesting the token is still invalid and error monitoring is compromised; this has been ongoing since yesterday.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a recurring issue.
The E2E tests are still failing, indicating ongoing issues with the platform's core functionality.
Spanner CPU usage is at 60%, which is at the warning threshold. This should be monitored closely for further increases.
The Platform Activity query is still failing, preventing monitoring of key business metrics. This is the second consecutive failure.
The Sentry API error (403) continues, suggesting the token is still invalid and impacting error monitoring capabilities.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are failing, indicating potential issues with core functionality.
The overall error rate in Cloud Run Metrics is high (8.6%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a persistent issue.
The Platform Activity query continues to fail, preventing monitoring of key business metrics; this needs to be resolved to track post creation and moderation queue status.
The Sentry API is still returning a 403 error, indicating a problem with the token; this is preventing proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services; this is a recurring issue.
The overall error rate in Cloud Run Metrics remains high (7.3%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The E2E tests are still failing, indicating ongoing issues with application stability.
The Platform Activity query failure persists, preventing monitoring of key business metrics.
The Sentry API error (403) continues, indicating a problem with the Sentry token which impacts error monitoring capabilities.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are failing, indicating a potential regression in functionality.
The overall error rate in Cloud Run Metrics is high (8.3%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query failed again; this is the second consecutive failure and prevents monitoring of key business metrics.
The Sentry API is returning a 403 error, indicating a problem with the token; this needs to be addressed to ensure proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services; this is a recurring issue.
The overall error rate in Cloud Run Metrics remains high (8.3%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The E2E tests are still failing, indicating a potential regression in the codebase.
The Platform Activity query continues to fail, preventing monitoring of key business metrics. This needs immediate attention.
The Sentry API is still returning a 403 error, indicating that the token is likely invalid or lacks the necessary permissions. This is preventing proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are failing, indicating potential functional issues within the platform.
The overall error rate in Cloud Run Metrics remains high (8.2%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a persistent issue.
The Platform Activity query continues to fail, preventing monitoring of key business metrics. This needs to be resolved to track post creation and moderation queue status.
The Sentry API is still returning a 403 error, indicating a problem with the token. This is preventing proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services. This is a recurring issue.
The overall error rate in Cloud Run Metrics remains high (8.5%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The E2E tests are failing, indicating a potential regression in functionality.
The Platform Activity query is failing again, preventing monitoring of key business metrics, and needs investigation.
The Sentry API is returning a 403 error, indicating a problem with the token, which impacts error monitoring and has been ongoing since yesterday.
Cloud Run Health checks are failing for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, which correlates with high error rates in Cloud Run Metrics, continuing a trend from previous runs.
The overall error rate in Cloud Run Metrics is high (8.1%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a recurring pattern.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics. This is the second consecutive failure.
The Sentry API error (403) continues, suggesting the Sentry token is still invalid and needs to be replaced to restore error monitoring capabilities.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are still failing, indicating ongoing issues with core functionality.
The overall error rate in Cloud Run Metrics remains high (8.0%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query continues to fail, preventing monitoring of key business metrics; this is the second consecutive failure.
The Sentry API is still returning a 403 error, suggesting the token issue has not been resolved and error monitoring is compromised.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
The overall error rate in Cloud Run Metrics remains high (7.0%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This is a persistent issue.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics. This is the second consecutive failure.
The Sentry API is still returning a 403 error, indicating a problem with the token. This is preventing proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services. This is a recurring issue.
The overall error rate in Cloud Run Metrics is high (7.5%), driven by errors in selio-vibe-crawler and other services experiencing health check failures. This has been a recurring issue.
The E2E tests are still failing, indicating a regression in functionality. This is the second consecutive failure.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics. This is the second consecutive failure.
The Sentry API is still returning a 403 error, indicating a problem with the Sentry token that needs to be addressed to restore error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services, indicating a persistent underlying issue.
E2E tests are failing, indicating issues with core functionality.
The overall error rate in Cloud Run Metrics remains high (7.7%), driven by errors in selio-vibe-crawler and other services experiencing health check failures.
The Platform Activity query is failing again, preventing monitoring of key business metrics. This needs to be resolved to track post creation, moderation queue, etc.
The Sentry API is still returning a 403 error, indicating a problem with the token. This is preventing proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-crawler-agents, and selio-vibe-scale-miner, correlating with high error rates in Cloud Run Metrics for these services. This has been ongoing.
Spanner CPU usage is at 52%, which is within normal limits, but should be monitored for trends as it's relatively high.
The E2E tests are still failing, indicating a potential regression in the codebase.
Platform Activity query failed again, preventing monitoring of key business metrics. This needs to be resolved to track post creation and moderation queue status.
The Sentry API is still returning a 403 error, indicating a potential issue with the Sentry token. This prevents proper error monitoring.
Cloud Run Health check failures persist for selio-vibe-conversation-analyser, selio-vibe-scale-miner and selio-vibe-crawler-agents, correlating with high error rates in Cloud Run Metrics. This is a continuation of yesterday's critical alerts.
E2E tests are failing, indicating potential issues with core functionality.
The overall error rate in Cloud Run Metrics is high (9.3%), driven by errors in selio-vibe-crawler and other services experiencing health check failures, continuing a trend from previous runs.
Platform Activity query failed again; this needs investigation as it blocks monitoring of key business metrics.
The Sentry API is returning a 403 error, indicating a problem with the token; this is preventing proper error monitoring.
Score Formula (recalculated nightly)
score = base(raw:0.2, processed:0.5, skill:0.8) + usage(min(used*0.05, 0.15)) + fresh(24h:+0.05, 7d:+0.02) + human(+0.1) - age_decay
Memories are never deleted. They sink in ranking but stay forever. Raw → Processed (5+ uses) → Skill (15+ uses + human validated).