This page describes what the hosted Tupshar service exposes. Detailed operational dashboards and internal telemetry are managed by Upside Down Research and are not publicly accessible.

Health Endpoints

Two health endpoints are available on the API host:

Liveness — always returns 200 while the process is running:

GET /healthz
→ 200 {"status":"ok"}

Readiness — returns 503 until schema bootstrap completes at startup, then 200 thereafter (it does not re-probe live DB connectivity on each call):

GET /readyz
→ 200 {"status":"ready"}   (DB connected)
→ 503 {"status":"starting"} (still bootstrapping at startup)

Prometheus Metrics

The service exposes Prometheus metrics on a separate port 9090 at /metrics. This port is not exposed publicly; scraping is internal to UDR infrastructure.

All metrics carry the tupshar_ prefix. The following metrics are emitted:

Metric	Type	Description
`tupshar_http_requests_total`	Counter	Total HTTP requests by method, path, and status
`tupshar_http_request_duration_seconds`	Histogram	HTTP request latency
`tupshar_auth_verify_total`	Counter	Total authentication verifications
`tupshar_auth_cache_hits_total`	Counter	Auth token cache hits
`tupshar_auth_cache_misses_total`	Counter	Auth token cache misses
`tupshar_ft_query_duration_seconds`	Histogram	Full-text query execution latency
`tupshar_tokenize_duration_seconds`	Histogram	Tokenisation latency
`tupshar_ratelimit_throttled_total`	Counter	Requests rejected by rate limiter

If you observe unexpected behaviour (sustained errors, slowness), contact paul@upside-down-research.com. We do not offer customer-facing access to metrics or dashboards during the research preview.