Service Level Agreement
Our reliability targets, how we measure them, and what we commit to
Last updated 2 June 2026. This page is deliberately conservative. We publish targets we can stand behind and give you a live status endpoint to check us against, rather than quote an annual uptime figure we do not yet have the track record to prove.
What this covers
Two distinct things: the availability of the API service itself, and the freshness of the feed data behind it. It does not cover the availability of the upstream government agency feeds we aggregate, which are outside our control. We detect and surface those outages rather than hide them.
Service availability
Target: 99.9% monthly availabilityof the API service. The API is served from Vercel's edge network with data from Supabase, both SOC 2 Type II infrastructure. It keeps serving the last known data even if our poller is offline, so a feed problem does not take the API down.
Honest caveat: our verifiable operating record began in April 2026. We report measured uptime as it accrues rather than guarantee an annual figure until a full 12-month track record exists. Availability (can you reach the API and get a response) is separate from freshness (how recently each feed last updated).
How we measure it
- 28 per-feed uptime monitors with alerting.
- A public /api/v1/status endpoint, no key required, that reports live feed health. Check it yourself, any time.
- Platform monitoring on Vercel for the API layer.
Feed freshness and self-healing
Each feed is re-polled on an adaptive interval, typically every 1 to 5 minutes. The ingestion engine is built to recover from failure without us touching it:
- Unhealthy containers are detected and restarted automatically (every 30 seconds).
- The poller fails its own healthcheck if it stops writing data, which triggers a restart.
- A per-feed circuit breaker quarantines a misbehaving upstream feed instead of letting it stall the rest.
- New builds deploy and roll over automatically.
This is a genuine strength: the system is designed to keep running, and to recover on its own when something upstream breaks.
Upstream dependencies
We aggregate official government emergency feeds across all states and territories. When an agency feed goes down or changes, that is outside our control. Our commitment is to detect it, keep serving the last known good data, and reflect the feed's status honestly on the status endpoint rather than present stale data as current.
Incident response
We monitor continuously and respond to service incidents as a priority. For confirmed outages affecting paid customers, we notify by email and post status to the public status endpoint. Security issues follow the disclosure process on our Security page.
Support response by tier
| Plan | Support | First-response target |
|---|---|---|
| Free | Community and email | Best effort |
| Starter | Best effort | |
| Developer | 2 business days | |
| Enterprise | Dedicated | Per contract |
Business day means Australian Central Time business hours. We do not offer a 24/7 support line.
Custom SLA for Enterprise
Enterprise customers can negotiate a contractual SLA with specific availability targets, response times, and service credits, agreed per deal. If you have procurement requirements, get in touch and we will scope them with you.