Status
Service status
Two live checks of the gateway, taken from the Serviqo web server. Everything below them is labelled with what it is: a measurement, or configuration.
Gateway operational
The gateway answered the liveness check in 76 ms. It reports its database as connected.
- Gateway response
- 76 ms
- Serviqo server to gateway, liveness check. Excludes model time.
- Gateway database
- Connected
- What the gateway itself reports on its readiness check. Every key lookup depends on it.
- Last checked
- Readings are reused for up to 15s; the page refreshes every 30s.
Gateway operational. The gateway answered the liveness check in 76 ms. It reports its database as connected.
Model tiersLink to this section
The five tiers the gateway is configured to serve, with the limits each one enforces.
| Model | What it is for | Context | Max output |
|---|---|---|---|
| serviqo-t1Flash | Cheap and quick. Good for small edits, questions and iteration. | 131,072 | 8,192 |
| serviqo-t2Coder | Purpose-built for code. The sensible default for agent work. | 262,144 | 16,384 |
| serviqo-t3Plus | Strong generalist with a very large context window. | 262,144 | 32,768 |
| serviqo-t4Pro | High-capability coding and long-horizon agent work. | 262,144 | 32,768 |
| serviqo-t5Flagship | The heavy one. Reach for it when the others stall. | 262,144 | 65,536 |
Context and output limits are in tokens, and are the caps the gateway enforces. This table is read from configuration — it tells you what the gateway is set up to serve. It is not a per-tier health check, and no row here has been contacted to draw this page. See what is measured.
What this page measuresLink to this section
The distinction matters more on a status page than anywhere else on this site, so it is written out rather than implied by a colour.
- Measured, live
- Two unauthenticated requests to the gateway, sent together from the Serviqo web server.
GET /health/livelinessis timed end to end: a success means the gateway process is up and answering on that path.GET /health/readinessis read, not timed: the gateway answers it only after a round trip to its own database and reports the result, and every key lookup depends on that database. Together they are reused for up to 15 seconds so that a room full of open tabs, or one impatient refresh key, costs the gateway a single pair of requests; the timestamp beside them always tells you how old they are. - Measured, but narrower than it looks
- The response time is the round trip from this server to the gateway on the internal network. It is not what your request will feel like. A completion takes as long as the model takes, plus your own network, and neither of those is observed here. The database line is the gateway’s own statement about itself, relayed as given; this page adds nothing to it.
- Reported from configuration, not measured
- The tier list, the context windows and the output caps. They are read from this site’s own configuration, which mirrors the gateway’s and is kept in step with it by hand — so if the two ever drift, this table is the one that is wrong. Nothing is sent to a tier to produce it, and it cannot tell you that a given tier answered a moment ago.
- Not measured at all
- Per-tier availability. Checking a tier means paying for a completion on it, on every load of this page, and we are not going to bill that to make a status page look busier. The gateway’s own per-model check exists but needs a key, and this site holds none. If you hold one there is a live check that costs nothing and needs no privileged access: the key checker asks the gateway what your key may call, and the answer comes from the gateway rather than from us.
- Deliberately absent
- An uptime percentage and an incident history. This site keeps no database and stores nothing, so there is no record to compute them from. Publishing either would mean inventing it.
If this page and your editor disagree, believe your editor
A green banner here means the gateway answered its health checks from inside our network a few seconds ago. It is a useful signal and a narrow one. If your requests are failing while this reads operational, the problem is further along than these checks can see — start with your key and the base URL in the setup guide.