Skip to main content

Audit passes against forecasts

A pass audit lines up every pass over a past window with the contacts you actually flew: what each pass could have delivered, what it did deliver, and which ones were lost. It turns "the ground network felt unreliable last month" into a number per station.

What you send​

A CSV of flown passes, posted to POST /telemetry/pass-logs, with these columns in any order:

satellite_id,antenna_id,aos_utc,los_utc,bytes_delivered,status

An audit needs your fleet first: every row's antenna and satellite must be in your network config, and a log with a bad row writes nothing. On an account with no network config, the upload answers 404 {"error": "no_network_config"}, and the audit under Past on Passes offers Load the demo fleet or Start from the catalog before it takes a log. A refused log lists its bad rows grouped by problem, each with a count and a few of the IDs it named.

A sample pass log​

On the demo fleet, the audit under Past on Passes offers Use a sample pass log. It uploads a log of the past week for two demo satellites, NOAA 21 on Ka band and Sentinel-2C on X band, with one or more rows inside each pass the audit of that week finds. The passes rain puts at risk failed, the ones whose antenna was still on another pass were missed, and of the rest most delivered, a few short and a few never logged, so the ledger shows every outcome. The API equivalent is GET /telemetry/pass-logs/demo-sample. The demo fleet's Hugging Face dataset carries a fixed-week pass-log-sample.csv in the same columns: audit it over the week it covers, which its dataset page names, since an audit of other days finds none of its contacts.

What you get​

The audit-ledger workflow template runs over the past window:

OutputMeaning
possible_gbWhat the pass could have delivered, counted at full lock where your telemetry shows the link lost lock. The totals count a pass at least at what it delivered, so delivered never reads above 100%
delivered_gbWhat your log says it delivered
delivered_exceeds_capacitytrue when your log says the pass delivered more than its link can carry at its best modulation for the whole pass, at full lock. The log is kept as logged; the log or the link's RF inputs may be off. The Platform marks the row "logged above link capacity". Absent on a pass within capacity
delivered_capacity_ratioWith delivered_exceeds_capacity, delivered over that ceiling, such as 1.5
statusThe pass's outcome: delivered, failed, missed or unlogged. A pass with nothing logged whose telemetry never showed lock is missed, and loses all its possible GB
totalsSums per ground station and overall. An antenna tracks one pass at a time, so passes that overlap on it count its best schedule once: possible GB totals can be less than the sum of the passes
unmatched_logsLogged contacts that matched no computed pass, each with its reason: no_pass_on_link (the run has no pass on that antenna and satellite: a satellite it did not cover, or a link never up) or between_passes (none at that time)
logged_contactsHow many logged contacts the audit read in its window. When it is 0, or none matched, the report says why instead of showing 0% delivered

Run it with POST /predictions, a past from and to, and dataset: audit. See Running and reading a workflow. In Platform, the audit under Past on Passes creates the audit-ledger workflow for you when you have none. On Explore with the demo fleet, which holds more satellites than an Explore run covers, the page limits the workflow to the satellites your pass log names before it runs.

Plans​

Operator and above audit with no time limit. Explore includes one free audit:

LimitExploreRefused with
Time48 hours from the account's first audit run403 free_audit_ended, on an audit run after the 48 hours. Reports already run stay readable.
Pass logLog up to 2 of your own satellites; any run covers at most 7 daysNever refused for the plan
SatellitesAn audit run covers at most 2 of your own satellites403 plan_required, with feature: "satellites"
RunsTest runs (deliver: false)

The 48 hours start when the first audit run starts, not at sign-up or at the first upload. After they end, an audit run answers 403 {"error": "free_audit_ended", "ended_at": "...", "plan": "standard"}, and the reports already run stay readable. GET /account/state reports the window as audit_trial: {"started_at": null, "ends_at": null, "active": true} before the first run, then both times, with active false once it has ended. On Operator and above, audit_trial is null. The audit under Past on Passes shows the time left and, once it ends, the upgrade to Operator.

Common questions​

What columns does a pass log need?

satellite_id, antenna_id, aos_utc, los_utc, bytes_delivered and status, in any order. The IDs must be in your network config, so load or build your fleet first. On the demo fleet, the audit in the Platform can upload a sample pass log for you.

What if a logged contact matches no pass?

It is listed in unmatched_logs, so you can check for a wrong ID, a wrong time, or a pass your network config does not describe.

Which plan includes pass audits?

Every plan. Explore includes one free audit: 48 hours from its first audit run, with each run covering up to 2 satellites. Operator and above audit with no time limit.

Next: run a free audit of your own passes, or try it on the demo fleet with its sample pass log.