OutageAS2 gateway outage
Aug 20, 2026 at 13:29 EDT to 10:07 EDT · 21 h · AS2 / HTTPS
Reconstructed from customer notification email.
Incident report ODT, 39 KB
Notice
Planned maintenance
Sep 22, 2026 at 18:00 EDT to Sep 23, 2026 at 02:00 EDT · 8.0 h · AS2 / HTTPS, FTP / SFTP, Customer Portal
Scheduled maintenance of production systems. AS2 / HTTPS, FTP / SFTP and the Customer Portal may be unavailable during this window.
Aug 20, 2026 at 13:29 EDT to 10:07 EDT · 21 h · AS2 / HTTPS
Reconstructed from customer notification email.
Incident report ODT, 39 KB
Aug 10, 2026 at 17:30 EDT to 18:14 EDT · 44 min · FTP / SFTP
Detected by automated monitoring. Cause and impact detail follow once reviewed.
Aug 10, 2026 at 15:04 EDT to 17:34 EDT · 2.5 h · AS2 / HTTPS
Detected by automated monitoring. Cause and impact detail follow once reviewed.
Published with its real duration. Not counted as downtime: the service was answering.
Aug 10, 2026 at 11:47 EDT to 18:17 EDT · 6.5 h · AS2 / HTTPS, FTP / SFTP
Reconstructed from customer notification email.
Incident report ODT, 37 KB
Aug 1, 2026 at 04:19 EDT to 04:24 EDT · 5 min · AS2 / HTTPS
Detected by automated monitoring. Cause and impact detail follow once reviewed.
Published with its real duration. Not counted as downtime: the service was answering.
Jul 21, 2026 at 10:16 EDT to 10:42 EDT · 26 min · Customer Portal
Detected by automated monitoring. Cause and impact detail follow once reviewed.
Jul 21, 2026 at 10:16 EDT to 10:42 EDT · 26 min · AS2 / HTTPS
Detected by automated monitoring. Cause and impact detail follow once reviewed.
Uptime covers a rolling 90 days. All times are US Eastern, and days break on Eastern midnight.
Overall is the unweighted mean of the services above, each measured against its own window — not a combined total. Its bar shows a split mark on days when some services were affected and others were not.
Announced maintenance is excluded from the measurement window, using the time it actually took rather than the window it was booked for.
Degraded periods are published with their real duration but are not counted as downtime, because the service was answering. Degraded hours are shown separately rather than hidden inside a flattering percentage.
Automated checks are one input, not the arbiter. Where a check and a recorded incident disagree, neither overrides the other: if any source says down, the service is down. Interruptions shorter than five minutes are recorded but not listed.
Machine-readable: status.json · feed.xml (events only). status.json publishes generated_at and expected_interval_s, so you can detect staleness without guessing our schedule.