Summary

  • Australia’s outage registers require carriers to publish a standardised, downloadable history of major and significant local outages, with entries generally due within three days of restoration.
  • The rules improve the evidence available after an outage, but each carrier reports from its own network position; the same underlying incident may appear in several registers.

The service is back. The historical record may still be three days away.

Since 30 June 2026, Australian carriers have had to maintain public registers of major outages and significant local outages affecting the networks they use to supply end users. The first records reach back to qualifying outages resolved from 31 March, including some that began earlier. A register must be prominent, accessible, kept electronically and downloadable as a structured CSV. It is designed to preserve information after the live incident has ended.

That timing matters. Live notices tell customers that a network is impaired and may change as engineers isolate a fault. The register serves a different purpose: it gives customers, regulators and researchers a record to inspect later. An entry is generally due no later than three days after the outage is resolved. If a carrier later discovers that the information is inaccurate or out of date, the register must be corrected within three days. Entries remain for at least two years after restoration.

The scope is narrower than “every outage.” A significant local outage must be an unplanned network impact that prevents ordinary use of a relevant carriage service, affects at least 1,000 services in regional Australia or 250 in remote Australia, and lasts or is expected to last more than six hours regionally or three hours remotely. Major outages are a separate category. Smaller interruptions and events outside those definitions are not automatically covered by this particular register duty.

The prescribed fields make comparisons possible: a carrier-assigned reference, outage type, start and end times, duration, affected places and service types, counts where known, and a cause category where disclosure is appropriate. Yet the system is not one national incident ledger. ACMA says downstream carriers may also have to publish records, so the same underlying outage can appear in multiple registers, each describing its effect on a different part of the supply chain. The reference is assigned by the carrier; the rules do not prescribe a regulator-issued identifier that joins all carriers’ entries.

That distinction is useful. A fibre break, power failure or system fault can disrupt wholesale and retail services differently. Multiple records can expose those different impacts rather than collapse them into a single “outage” total. But anyone counting incidents or affected services across downloadable files will need to detect overlaps. A common CSV structure helps; it does not itself deduplicate events.

The rules also recognise that no carrier sees everything. Some service counts are required only to the extent technically practicable and relevant. A cause can be withheld where the carrier reasonably believes disclosure could prejudice network or national security. A non-carrier retail provider must prominently link to the register of the carrier whose network it uses; it does not maintain a separate register under this part of the standard. These are design choices about visibility, not evidence that an omitted field was deliberately concealed.

Nor is a register entry a customer’s account of the outage. It records a network event in plain English, but its coverage and counts are bounded by what the carrier knows and by the standard’s thresholds. Restoration means the affected services have been restored as relevant to the event; it does not establish when every individual subscriber regained service, whether performance returned to normal, or whether anyone received compensation. Separate live communications duties continue during the incident.

The published rules create a valuable starting point for retrospective scrutiny. They do not yet demonstrate a national reliability trend. ACMA’s April–June 2026 report notes that one compliance assessment concerning the broader outage-communications standard was completed; it does not report a nationwide audit of register completeness or establish industry-wide compliance. The next evidentiary step is to compare actual registers: publication delays, correction histories, missing fields, overlapping references and how consistently downstream providers link to the relevant network record.

Sources: ACMA’s outage-register rules; current CCO Standard, Part 4; ACMA’s commencement notice; industry guidance; ACMA’s April–June 2026 report.

Sources