32 docs tagged with "operations"

How to prove a bot-initiated BYE or transfer made it end to end from the voice PBX through the SIP edge to the carrier, and which alerts to build on that chain.

TelAPI Fastify HTTP/WebSocket API behind a managed load balancer — operations, runbook, configuration, troubleshooting.

What TelSys, TelPro, and RTPEngine log during a call, how log levels are used, and how to correlate application logs with SIP signaling.

One summary log record per call — timing, bot context, provider identifiers, latency, error/fallback, and correlation IDs — queryable in SigNoz and persisted to AppLog.

How a call enters Asterisk Stasis, how TelPhi drives it from StasisStart to teardown, and how the max-call-duration limit is armed and enforced.

Back up and restore Delphi deployment configuration across bootstrap env, generated config bundles, AWS SSM, Secrets Manager, and Postgres-backed settings.

Migration vs runtime Postgres users — MIGRATION_DATABASE_URL for schema work, DATABASE_URL for application CRUD, and optional DATABASE_SECRET_ARN for day-2 rotation.

PostgreSQL 17, Redis 7, and PgBouncer with attached block storage — operations, runbook, configuration, troubleshooting.

Operator-side symptoms — container, network, secret, migration — with pointers into the canonical per-service runbooks.

The two scripts every service uses — init.sh brings the service up from a clean shell, update.sh wraps it with backups and confirmation for steady-state operations.

Shared SSH, Docker Compose, healthcheck, log, environment, network, and OTel collector checks for Delphi service hosts.

How Delphi expects SIPp / Vodafone load tests to be run, certified single-voice capacity, and STT/TTS WebSocket resilience under load.

Private HTTPS TTS media cache (Caddy + media-upload) at static private IP 10.0.1.30 — operations, runbook, configuration, troubleshooting.

How Delphi services send logs, traces, metrics, health signals, and call debugging data through OTel collectors into SigNoz.

Per-service operations index — purpose, runtime, ports, dependencies, and links to each service's operations page.

Scaler and Tasker — autoscaling, background jobs, cron, DB backups, and email notifications.

Operator playbooks for restoring a broken Delphi environment. Conservative defaults, clear "stop and ask" points, and links to the canonical runbooks.

Re-run init / update on a Delphi service host cleanly, including the case where the previous update failed mid-way (partial migration, half-pulled image, half-applied config).

When the Ops service is up but a queue has stalled — diagnose, drain, and re-enqueue safely.

Which Delphi services must be restarted after image, config, SSM, Secrets Manager, TLS, and database changes.

Reference for the S3 objects Delphi stores for a recorded conversation, including audio tracks, transcripts, turn artefacts, and temporary source files.

What each Delphi service logs, where the logs land (SigNoz via OTLP and stdout filelog), and the full TelWeb and TelPhi log inventories — server-side structured logs, call/telephony events, CDR, AI/DB spans, audit/auth events, route-handler logs, and client-side console output.

SigNoz observability backend on static private IP 10.0.1.10 — Foundry deploy, version upgrades, legacy migration, alert sync, and troubleshooting.

What Kamailio and TelSys log for SIP — capture points, methods (OPTIONS, CANCEL, PRACK), response codes (481, 487), and how records reach the Debug SIP ladder.

Catalog of SIP trunk health logs emitted by TelPhi and TelAPI — when each fires, what it means, and which payload fields appear.

Configure, operate, and verify seven.io SMS delivery through the Tasker worker.

HTTP/HTTPS proxy for private-network instances — operations, runbook, configuration, troubleshooting.

SIP proxy (Kamailio), media proxy (RTPEngine), WebRTC gateway (Janus), and TURN server — operations, runbook, configuration, troubleshooting.

How super-admins search retained SigNoz telemetry by trace ID, call ID, or SIP header value from the TelWeb Trace Debug view — friendly header mode and raw ClickHouse SQL mode.

How the TOBi transactionId propagates from the managed LLM provider through TelAPI, Redis, Kamailio, TelSys, and OTel spans so every event in a conversation can be correlated under one stable id.

Asterisk (TelSys) plus the AI conversation engine (TelPhi) — operations, runbook, configuration, troubleshooting.

TelWeb management dashboard behind Caddy — operations, runbook, configuration, and troubleshooting.