Database
PostgreSQL — links, accounts, analytics and everything else that persists.
- Result
- Responding to queries.
- Latency
- 22 ms
Status
All checks are passing
The probes below ran when you loaded this page, and no component is currently flagged by an operator.
All systems operational
Checked
Live probes
Measured when you opened this page
Not a cached summary and not a scheduled sample — these two checks ran against the running system as this page was rendered.
Passing
Cache
No Redis configured; the in-process fallback is handling cache and rate limits.
- Result
- In-process fallback. Correct on a single instance, not shared across replicas.
- Latency
- Not measured
The cache is running on the in-process fallback
No Redis URL is configured, so caching and rate limiting use an in-memory store inside each application process. That is correct behaviour on a single instance and incorrect across several: two instances will not share a rate-limit window. This distinction matters during an incident, so it is stated rather than hidden behind a green tick.
Components
Operator-reported state
Set by whoever is handling the system. A component can be flagged here even while both probes above pass — for example when a downstream provider is failing.
Website
Marketing site and documentation
OperationalDashboard
Link management and settings
OperationalRedirect service
Short link resolution
OperationalAPI
Public REST API
OperationalAnalytics
Click recording and reporting
Operational
Incidents
Recent incidents
Written by the person handling the incident. Newest first; an entry with a resolved time is closed.
No incidents recorded
Nothing has been written to the incident log. When something goes wrong, the update and its resolution appear here.
How to read this page
Why there is no uptime percentage
Most status pages open with a figure like 99.98% and a bar of ninety coloured squares.
Nothing here samples availability on a schedule
This page reports two things and nothing more: probes that ran the moment you loaded it, and states an operator has set by hand. There is no external monitor recording minute-by-minute availability, so an uptime percentage or a 90-day history strip would be a number we made up — and during an incident that is precisely the number people rely on.
If you need contractual availability reporting, an independent monitor pointed at your own links is the honest way to get it. Ask us at [email protected] and we will tell you exactly what we do and do not measure.
Planned maintenance is published as a component in the Maintenance state with an incident entry describing the window, before the work starts rather than after someone notices.