MandiIQ Public status
resolving instance builds unchecked no checks yet not checked yet

Which instance is answering, and
why the newest price is -

MandiIQ runs the same image behind two hostnames, so one of them can be down while the product keeps working. This page probes both directly and reports what each one says about itself, then explains the gap between the newest price it holds and today. Everything here is read from /health on the live services - nothing is cached, and no reason is listed that the service did not report.

probing

Contacting both instances...

Downtime timeline

One entry per transition between answering /health and not answering it, written by the 60-second check the shared page shell (site.js) runs on every page. This is the long form of the record the other pages show in short: this is a page, not a monitor, an outage that begins and ends while no tab is open is invisible, an entry can never be more precise than the 60-second check, and the reason shown is the text the instance or the platform edge returned - never a guess about it. Entries are kept in this browser only, capped to the last few outages per instance, and clearing site data erases them. The pill in the header is the one-line form of this list.

No checks recorded yet.

Build skew

What each instance says about its own build, taken only from the fields its /health payload actually publishes. A field an image does not send is absent - not zero, not false, and not a failure - so an older build is named as one here instead of being read as a broken one below.

Comparing published payloads...

Instances side by side

Both columns answer the same image behind two hostnames, and a value marked in rust differs between the two. A row reading not reported is a field that build does not publish at all; the panel above says which build that is.

Loading...

Why the data is behind

Derived from the serving instance's /health payload, plus the two upstream findings recorded in NORTHFLANK_DEPLOY.md.

Documented routes

The routes this build is expected to serve, probed against the instance that is answering. A 404 in this table means the deployed build is older than the repository - not that the route failed. Probed every five minutes, not on every refresh, because some of these are real computations.

Loading...

Raw payloads

The unedited /health response from each instance, so nothing above has to be taken on trust.