System status

Live status, measured availability, incident history, release log and service levels for ArkCloud Health.

All systems operational

Last automated check 2026-11-09 00:43 UTC. Measured history begins 2026-10-02 13:13 UTC.

Updated 2026-11-09 00:43 UTC ·

Components

Measured from automated checks about once a minute. Each bar is one day (UTC), oldest on the left. Availability counts only minutes that were checked; monitored is the share of minutes that were checked. Grey means no data, never "up". Blue marks a day whose only outage minutes came while a recorded release was being installed (the service restarts); those minutes still count against availability, and any other outage minute makes the day red.

Public website & edge network
Requests the health endpoint through the public address (arkcloudhealth.com), via the edge network, the way a customer's browser does. Catches failures that the server cannot see from the inside.
Operational
Last 24 hours
99.650% · 99.1% monitored
7 UTC days
99.741% · 84.4% monitored
30 UTC days
95.28% · 90.7% monitored
90 UTC days
96.28% · 92.0% monitored
Application & API
Requests the application's health endpoint directly on the server. Operational = HTTP 200 within 5 seconds.
Operational
Last 24 hours
99.790% · 99.1% monitored
7 UTC days
99.932% · 84.4% monitored
30 UTC days
99.884% · 90.7% monitored
90 UTC days
99.829% · 92.0% monitored
Database
The application's own database check, as reported by its health endpoint.
Operational
Last 24 hours
100% · 98.9% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
100% · 90.6% monitored
90 UTC days
100% · 91.8% monitored
Cache & job queue
The application's own Redis check, as reported by its health endpoint.
Operational
Last 24 hours
100% · 98.9% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
100% · 90.6% monitored
90 UTC days
100% · 91.8% monitored
Background jobs
The background worker's heartbeat, as reported by the health endpoint. Covers reminders, statements, remittance processing and other scheduled work.
Operational
Last 24 hours
100% · 98.9% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
100% · 90.6% monitored
90 UTC days
99.994% · 91.8% monitored
AI assistant
Whether the AI model answered the application's most recent requests (real ones, or a one-word test request sent every few minutes when it is idle), as reported by the health endpoint. Degraded = the last request failed; down = the last two failed. AI features say when they are unavailable and the rest of the application is unaffected.
Operational
Last 24 hours
100% · 98.9% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
90.06% · 82.8% monitored
90 UTC days
90.06% · 64.2% monitored
Search
The search engine's own health endpoint. If search is down, the application falls back to slower database search.
Operational
Last 24 hours
100% · 99.1% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
100% · 90.7% monitored
90 UTC days
99.972% · 92.0% monitored
Database backups
Age of the newest verified database backup on the server. Operational = less than 26 hours old (backups run daily); degraded = 26-50 hours; down = older or missing.
Operational(newest backup 15 h old)
Last 24 hours
100% · 99.1% monitored
7 UTC days
100% · 84.4% monitored
30 UTC days
100% · 90.7% monitored
90 UTC days
100% · 92.0% monitored

operational
operational, partly monitored
degraded
restarted for a release
outage recorded
no data

Integrations

Optional services that depend on configuration for this deployment. 'Not enabled' means the feature is switched off, not that it has failed. Enabled integrations are not yet health-checked.

Email deliveryEnabled (not monitored)
Text messages (SMS)Enabled (not monitored)
Clearinghouse (claims and remittances)Not enabled
Card paymentsNot enabled

Incident history (90 days)

Every incident that affected the service, with its cause and what was changed. Updates in a timeline are never edited or deleted; a correction is a new update.

Short outages while build jobs ran on the production server

INC-2026-1101-01 · Resolved · Major impact (SEV1)

Started: 2026-11-01 02:07 UTC
Resolved: 2026-11-01 04:53 UTC
Duration: 2 h 46 min (approximate, from logs)
Affected: Public website & edge network, Application & API

Between about 02:07 and 04:53 UTC on 1 November, the website and application stopped answering the uptime probe in four short periods. Build and test jobs running on the same server as production used enough memory and processor time to starve the live service.

Customer impact: Pages and sign-in may have failed or timed out for a few minutes at a time, around 02:08, 03:00, 03:48 and 04:42 UTC. The probe recorded 18 minutes of application downtime for the day (UTC). The exact per-request impact has not been established.

Root cause: Release builds and test runs share the production server. Extra type checks and test runs were started outside the server's one-heavy-job-at-a-time slot and ran alongside a release build and the test gate, starving the live service.

Corrective actions: Every heavy job - builds, test gates and type checks - now waits for the single heavy-job slot. Moving builds and test runs to a separate build host is in progress.

Resolved · 2026-11-01 04:53 UTC - Service recovered after each period; heavy jobs now run one at a time. Published retrospectively. (published 2026-11-01 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Automatic probe incidents SUP-005192, SUP-005193, SUP-005205, SUP-005226, SUP-005273 and SUP-005274 (probe down, two consecutive minutes each); the release log of build, test and type-check runs.

Public website served an error page while the service was running

INC-2026-1002-02 · Resolved · Critical impact (SEV1)

Started: 2026-10-02 03:27 UTC
Resolved: 2026-10-02 05:15 UTC
Duration: 1 h 49 min (approximate, from logs)
Affected: Public website & edge network

After the connector that links this server to the edge network was restarted, public traffic was routed to a second, unauthorised connector on another machine, which could not reach the service. Visitors were shown an error (503) page while the service itself was healthy.

Customer impact: Public access was interrupted for some or all of this window. The exact public impact window has not yet been established from edge logs; the times shown are bounded by host service events.

Root cause: A second connector on another machine had been using the same credential since 2026-09-20 and could not reach the service.

Corrective actions: The connector credential was rotated, locking the second connector out, and this server's connector was reconfigured with the new credential. Monitoring through the public edge (this page) now detects this failure mode.

Resolved · 2026-10-02 05:15 UTC - Edge connectivity restored with a rotated credential. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Windows service events: connector crash loop 03:27:04-03:43:14 UTC; replacement connector registered 05:15:41 UTC (edge connector list).

Release rolled back after a slow start

INC-2026-1002-01 · Resolved · Minor impact (SEV3)

Started: 2026-10-02 02:52 UTC
Resolved: 2026-10-02 03:04 UTC
Duration: 11 min (approximate, from logs)
Affected: Application & API

A release took longer to start than the health check allowed and was rolled back automatically, then installed successfully on a second attempt.

Customer impact: The service was interrupted for up to about 6 minutes during the first attempt and about 2 minutes during the second.

Root cause: The new build took about 50 seconds to start for the first time (most likely file scanning of the newly copied files), longer than the health-check window.

Corrective actions: Releases are now test-started before promotion (which warms them), and the health-check window was extended to 240 seconds.

Resolved · 2026-10-02 03:04 UTC - Installed successfully on the second attempt. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Deployment ledger: PROMOTION_FAILED 02:58:34Z, retry 03:01:58Z, PROMOTION_SUCCEEDED 03:04:05Z.

Service unavailable for 22 minutes after an interrupted release

INC-2026-0930-01 · Resolved · Critical impact (SEV1)

Started: 2026-09-30 13:53 UTC
Resolved: 2026-09-30 14:15 UTC
Duration: 22 min
Affected: Application & API, Public website & edge network, Background jobs

A release stopped the services and was interrupted while copying the new build, leaving an incomplete build installed and the services stopped.

Customer impact: The service was unavailable for 22 minutes (13:53-14:15 UTC). Background jobs were also stopped. No data was lost.

Root cause: The promotion script, on failure part-way through copying, recorded the failure but did not restore the previous build or restart the services.

Corrective actions: Restored by promoting the previous build at 14:15 UTC. Releases are now staged and test-started before promotion, and the promotion script takes an exclusive lock. Automatic restore on a mid-copy failure remains an open corrective action.

Resolved · 2026-09-30 14:15 UTC - Service restored by promoting the previous build. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Deployment ledger operation aa7856a0 (attempt with no outcome) and 21e14f74 (restore, succeeded 14:15:43Z); post-incident notes.

Release failed its health check and was rolled back

INC-2026-0927-01 · Resolved · Minor impact (SEV3)

Started: 2026-09-27 01:52 UTC
Resolved: 2026-09-27 01:56 UTC
Duration: 4 min (approximate, from logs)
Affected: Application & API

A new build did not pass the automatic health check after installation and was rolled back automatically.

Customer impact: The service was interrupted for up to about 4 minutes while the release was attempted and rolled back.

Root cause: The new build was unhealthy on start; the automatic health check caught it.

Corrective actions: Automatic rollback by the promotion script. The release was installed successfully at 02:12 UTC.

Resolved · 2026-09-27 01:56 UTC - The release was rolled back automatically. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Deployment ledger operation 65969d08 (PROMOTION_FAILED at HEALTH_CHECK, rolled back) and a42dbbb4 (succeeded).

Service unavailable after a release

INC-2026-0924-01 · Resolved · Critical impact (SEV1)

Started: 2026-09-24 09:57 UTC
Resolved: 2026-09-24 10:00 UTC
Duration: 3 min
Affected: Application & API, Public website & edge network

A release left both application services stopped and the installed build without its static files. The server returned no response.

Customer impact: The service was unavailable for about 3 minutes.

Root cause: The installed build was incomplete and the services were not restarted.

Corrective actions: Emergency rollback to the previous build (09:59:41-10:00:54 UTC); the release was re-installed successfully at 10:14 UTC.

Resolved · 2026-09-24 10:00 UTC - Resolved by rolling back to the previous build. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Deployment ledger operations 3505d743 (install), 02f3949c (EMERGENCY_ROLLBACK, reason recorded), 0066ce95 (re-install).

Site rendered without styles after a release

INC-2026-0923-01 · Resolved · Major impact (SEV2)

Started: 2026-09-23 23:13 UTC
Resolved: 2026-09-23 23:33 UTC
Duration: 19 min (approximate, from logs)
Affected: Application & API

A release installed an application build whose static files (stylesheets, fonts, scripts) were missing. Pages loaded without styling and did not become interactive.

Customer impact: Pages were reachable but unstyled and not interactive for about 19 minutes. No data was lost.

Root cause: The build was installed with no static/ directory; the promotion script did not verify that the build was complete before installing it.

Corrective actions: Emergency rollback to the previous build at 23:33 UTC. The promotion script now refuses an incomplete build (recorded refusals from 2026-09-25).

Resolved · 2026-09-23 23:33 UTC - Resolved by rolling back to the previous build. Published retrospectively. (published 2026-10-02 12:00 UTC)

Published retrospectively; customers were not notified at the time. Evidence: Deployment ledger entries 2026-09-23T23:13:56Z (install) and 23:33:02Z (EMERGENCY_ROLLBACK, reason recorded).

Releases and maintenance (90 days)

Installing a release stops and restarts the application, interrupting service for about 1 to 3 minutes. Every installation since 2026-09-24 is listed below from the deployment ledger, with its measured interruption. Measured

190 release installations in 90 days; total interruption at most 1 h 11 min.

Release installations
Started	Interruption (at most)	Outcome
2026-11-08 23:42 UTC	None	Completed
2026-11-08 20:57 UTC	None	Completed
2026-11-08 18:27 UTC	None	Completed
2026-11-08 16:15 UTC	None	Completed
2026-11-07 23:37 UTC	None	Completed
2026-11-07 19:38 UTC	None	Completed
2026-11-07 19:08 UTC	None	Completed
2026-11-07 18:05 UTC	None	Completed
2026-11-04 12:43 UTC	None	Completed
2026-11-03 06:48 UTC	None	Completed

Show all 190 installations

Earlier release installations
2026-11-03 03:21 UTC	None	Completed
2026-11-02 23:49 UTC	None	Completed
2026-11-02 16:40 UTC	None	Completed
2026-11-02 02:58 UTC	None	Completed
2026-11-02 01:13 UTC	None	Completed
2026-11-01 21:56 UTC	None	Completed
2026-11-01 20:56 UTC	None	Completed
2026-11-01 18:41 UTC	None	Completed
2026-11-01 18:05 UTC	None	Completed
2026-11-01 16:56 UTC	None	Completed
2026-11-01 15:24 UTC	None	Completed
2026-11-01 14:04 UTC	None	Completed
2026-11-01 10:57 UTC	None	Completed
2026-11-01 10:16 UTC	None	Completed
2026-11-01 09:18 UTC	None	Completed
2026-11-01 00:20 UTC	None	Completed
2026-10-31 23:43 UTC	None	Completed
2026-10-31 23:15 UTC	None	Completed
2026-10-31 22:52 UTC	None	Completed
2026-10-31 16:18 UTC	None	Completed
2026-10-31 15:55 UTC	None	Completed
2026-10-31 13:13 UTC	None	Completed
2026-10-31 12:36 UTC	None	Stopped before any change
2026-10-31 12:07 UTC	None	Completed
2026-10-31 11:41 UTC	None	Completed
2026-10-31 11:01 UTC	None	Completed
2026-10-31 01:50 UTC	None	Completed
2026-10-31 00:40 UTC	None	Completed
2026-10-31 00:36 UTC	None	Completed
2026-10-30 23:26 UTC	None	Completed
2026-10-30 22:25 UTC	None	Completed
2026-10-30 21:43 UTC	None	Completed
2026-10-30 21:15 UTC	None	Completed
2026-10-30 21:09 UTC	None	Completed
2026-10-30 20:38 UTC	None	Completed
2026-10-30 18:26 UTC	None	Completed
2026-10-30 18:04 UTC	None	Completed
2026-10-30 17:14 UTC	None	Completed
2026-10-30 16:38 UTC	None	Completed
2026-10-30 10:37 UTC	None	Completed
2026-10-30 09:12 UTC	None	Completed
2026-10-30 07:20 UTC	None	Completed
2026-10-30 06:04 UTC	None	Completed
2026-10-30 03:59 UTC	None	Stopped before any change
2026-10-30 02:09 UTC	None	Completed
2026-10-29 03:01 UTC	None	Completed
2026-10-28 09:31 UTC	None	Completed
2026-10-28 06:15 UTC	None	Completed
2026-10-28 05:26 UTC	None	Completed
2026-10-27 22:10 UTC	None	Completed
2026-10-27 19:14 UTC	None	Completed
2026-10-27 04:58 UTC	None	Completed
2026-10-27 02:02 UTC	None	Completed
2026-10-27 01:00 UTC	None	Completed
2026-10-26 22:57 UTC	None	Completed
2026-10-26 20:32 UTC	None	Completed
2026-10-26 20:12 UTC	None	Completed
2026-10-26 19:33 UTC	None	Completed
2026-10-26 17:20 UTC	None	Completed
2026-10-26 17:01 UTC	None	Completed
2026-10-26 15:42 UTC	None	Completed
2026-10-26 15:10 UTC	None	Completed
2026-10-26 14:40 UTC	None	Completed
2026-10-26 12:26 UTC	None	Completed
2026-10-26 00:21 UTC	None	Completed
2026-10-25 23:19 UTC	None	Completed
2026-10-25 21:28 UTC	None	Completed
2026-10-24 09:08 UTC	None	Completed
2026-10-23 14:41 UTC	None	Completed
2026-10-23 14:34 UTC	None	Completed
2026-10-23 14:22 UTC	None	Completed
2026-10-17 23:33 UTC	None	Completed
2026-10-17 22:53 UTC	None	Completed
2026-10-17 16:09 UTC	None	Completed
2026-10-17 15:03 UTC	None	Completed
2026-10-17 05:38 UTC	None	Completed
2026-10-17 05:11 UTC	None	Completed
2026-10-17 04:10 UTC	None	Completed
2026-10-15 18:38 UTC	None	Completed
2026-10-15 08:52 UTC	None	Completed
2026-10-14 22:11 UTC	None	Completed
2026-10-14 21:07 UTC	None	Completed
2026-10-13 14:00 UTC	None	Completed
2026-10-13 04:13 UTC	None	Completed
2026-10-12 23:49 UTC	None	Completed
2026-10-12 22:23 UTC	None	Completed
2026-10-12 03:15 UTC	None	Completed
2026-10-12 02:46 UTC	None	Completed
2026-10-12 01:42 UTC	None	Completed
2026-10-11 23:49 UTC	None	Completed
2026-10-11 23:45 UTC	None	Stopped before any change
2026-10-10 22:23 UTC	None	Completed
2026-10-10 19:36 UTC	None	Completed
2026-10-10 17:50 UTC	None	Completed
2026-10-10 14:36 UTC	None	Completed
2026-10-09 18:18 UTC	None	Completed
2026-10-09 17:41 UTC	None	Completed
2026-10-09 17:11 UTC	None	Completed
2026-10-09 14:49 UTC	None	Completed
2026-10-09 14:18 UTC	None	Completed
2026-10-09 13:58 UTC	None	Completed
2026-10-09 13:29 UTC	None	Completed
2026-10-09 11:55 UTC	None	Completed
2026-10-09 11:24 UTC	None	Completed
2026-10-09 02:32 UTC	None	Completed
2026-10-09 02:08 UTC	None	Completed
2026-10-08 21:59 UTC	None	Completed
2026-10-08 21:18 UTC	None	Completed
2026-10-08 20:18 UTC	None	Completed
2026-10-08 18:06 UTC	None	Completed
2026-10-08 14:03 UTC	None	Completed
2026-10-07 12:11 UTC	None	Completed
2026-10-07 10:08 UTC	None	Completed
2026-10-07 02:27 UTC	None	Completed
2026-10-06 20:10 UTC	None	Completed
2026-10-06 13:14 UTC	None	Completed
2026-10-06 06:09 UTC	None	Completed
2026-10-06 02:55 UTC	None	Completed
2026-10-06 01:15 UTC	None	Completed
2026-10-05 17:23 UTC	None	Completed
2026-10-05 16:08 UTC	None	Completed
2026-10-05 14:33 UTC	None	Completed
2026-10-05 13:02 UTC	None	Completed
2026-10-05 12:17 UTC	None	Completed
2026-10-05 11:38 UTC	None	Completed
2026-10-05 11:34 UTC	None	Completed
2026-10-05 11:34 UTC	1 min	Completed
2026-10-04 01:19 UTC	1 min	Completed
2026-10-04 00:27 UTC	1 min	Completed
2026-10-04 00:18 UTC	2 min	Completed
2026-10-03 22:31 UTC	1 min	Completed
2026-10-03 14:28 UTC	1 min	Completed
2026-10-03 12:59 UTC	1 min	Completed
2026-10-03 11:09 UTC	1 min	Completed
2026-10-03 10:27 UTC	1 min	Completed
2026-10-03 09:12 UTC	1 min	Completed
2026-10-03 08:45 UTC	1 min	Completed
2026-10-03 08:01 UTC	1 min	Completed
2026-10-03 00:29 UTC	1 min	Completed
2026-10-03 00:26 UTC	1 min	Stopped before any change
2026-10-03 00:22 UTC	1 min	Stopped before any change
2026-10-02 20:43 UTC	1 min	Completed
2026-10-02 19:50 UTC	1 min	Completed
2026-10-02 18:51 UTC	1 min	Completed
2026-10-02 18:23 UTC	1 min	Completed
2026-10-02 17:51 UTC	1 min	Completed
2026-10-02 17:25 UTC	1 min	Completed
2026-10-02 13:25 UTC	1 min	Completed
2026-10-02 13:11 UTC	1 min	Completed
2026-10-02 12:42 UTC	1 min	Completed
2026-10-02 11:50 UTC	1 min	Completed
2026-10-02 11:11 UTC	1 min	Completed
2026-10-02 03:01 UTC	2 min	Completed
2026-10-02 02:52 UTC	6 min	Rolled back automatically
2026-10-01 22:17 UTC	1 min	Completed
2026-10-01 18:50 UTC	1 min	Completed
2026-10-01 16:24 UTC	1 min	Completed
2026-10-01 15:49 UTC	1 min	Completed
2026-10-01 15:38 UTC	1 min	Completed
2026-10-01 11:15 UTC	1 min	Completed
2026-09-30 23:30 UTC	1 min	Completed
2026-09-30 22:03 UTC	1 min	Completed
2026-09-30 20:53 UTC	2 min	Completed
2026-09-30 14:14 UTC	1 min	Completed
2026-09-30 13:53 UTC	unknown	Interrupted - no outcome recorded
2026-09-29 23:30 UTC	None	Stopped before any change
2026-09-29 12:38 UTC	1 min	Completed
2026-09-29 11:53 UTC	None	Stopped before any change
2026-09-29 08:01 UTC	3 min	Completed
2026-09-27 15:27 UTC	1 min	Completed
2026-09-27 02:37 UTC	1 min	Completed
2026-09-27 02:11 UTC	1 min	Completed
2026-09-27 01:52 UTC	4 min	Rolled back automatically
2026-09-26 20:03 UTC	2 min	Completed
2026-09-25 15:09 UTC	1 min	Completed
2026-09-25 02:34 UTC	1 min	Completed
2026-09-24 16:11 UTC	1 min	Completed
2026-09-24 10:11 UTC	3 min	Completed
2026-09-24 09:59 UTC	1 min	Completed (emergency rollback)
2026-09-24 09:57 UTC	2 min	Completed

Planned work expected to interrupt service for more than 5 minutes will be announced on this page at least 48 hours in advance. Proposed - not yet adopted

Service levels

Every statement below shows its standing. Reviewed 2026-10-02.

Availability

No availability percentage (SLA) is published or guaranteed. Where a customer agreement sets an uptime term, that agreement governs. The availability figures on this page are measured, not promised.

Incident severity and response targets

Applies to service availability incidents. Proposed - not yet adopted These targets are not in effect until adopted in a written policy; they are published so performance can be measured against them.

Severity levels and proposed response targets
Severity	Definition	Acknowledge	First public update	Update cadence	Restore objective	Post-incident review
SEV1 Critical	The service is unavailable to customers, or there is a risk to the integrity of customer data.e.g. The website or application does not load; sign-in fails for everyone; data written incorrectly.	15 minutes	30 minutes	Every 60 minutes until resolved	4 hours	Published within 5 business days
SEV2 Major	A core function is unavailable or seriously degraded for some or all customers, with no reasonable workaround.e.g. Pages render without styles; scheduling or charting fails; background jobs stopped.	30 minutes	60 minutes	Every 2 hours until resolved	8 hours	Published within 10 business days
SEV3 Minor	Limited impact: a non-core feature fails, performance is reduced, or a brief interruption has a workaround.e.g. A release rolled back automatically within minutes; search slower than usual.	4 business hours	1 business day (if customers were affected)	On resolution	5 business days	Recorded internally; summary published

Measured performance (last 90 days)

Measured Computed from the incident records above, against the proposed targets.

Measured incident performance
Severity	Incidents	Median time to restore	Longest	Restored within objective	First update within target
SEV1	4	1 h 6 min	2 h 46 min	4 of 4	0 of 44 published retrospectively
SEV2	1	19 min	19 min	1 of 1	0 of 11 published retrospectively
SEV3	2	8 min	11 min	2 of 2	1 of 22 published retrospectively

Security incidents and breach notification Draft policy - not yet in effect

Source: Security Incident Response and Breach Notification Procedure (draft, not adopted)

Triage of a reported security incident: Within 4 hours of the report
Notice to an affected customer (covered entity) of a confirmed or reasonably suspected breach of PHI: Initial notice within 24 hours of discovery, as set out in the draft procedure and the business associate agreement template
Incident closed with root cause and corrective actions: Within 30 days of containment

Report a suspected security issue to [email protected].

Data protection and recovery Measured

Source: Contingency plan status (draft, 2026-10-01) and the backup checks on this page

Backup frequency: Daily at 02:30 US Pacific time (database dump, read back and checksummed after writing)
Backup retention: 14 days, on the same server
Backup location and encryption: Stored on the same disk as the database. Not encrypted. No off-site copy yet. These are open risks.
Recovery point (maximum data loss): Up to about 24 hours (time since the last daily backup)
Recovery time: Not yet measured for a full rebuild. Database restore alone: 3 seconds in the drill below.
Last restore test: 2026-10-01: a production backup restored completely into an isolated test database (all 345 tables, all 305 row-security policies). Not yet rehearsed into production.

How this page is produced

Each component is checked about once a minute by the platform's background worker, and every result is stored.
Availability = minutes the component answered (including degraded) / minutes we have a check for. A minute with no check is never counted as up.
Coverage = minutes with a check / minutes in the period. When the server itself is down, its checks cannot run, so low coverage can hide an outage; read the two together.
Monitoring runs on the same infrastructure it checks. There is no independent third-party monitor yet; the public-edge check goes out through the edge network and back.
Incidents before this page existed (before 2026-10-02) were reconstructed from the deployment ledger and host logs and published retrospectively; no customer was notified at the time.
All times are UTC. Durations from logs are marked approximate where the exact start or end was not observed.

This page reports operational status. It is not an audit report or a certification. No SOC 2 report is held, and no audit has been engaged yet. Our written security policies are drafts that are not yet in effect; see the HIPAA and Security pages.

Machine-readable: /api/status (JSON) and /status/feed.xml (Atom, incidents and maintenance).

All systems operational

Components

Public website & edge network

Application & API

Database

Cache & job queue

Background jobs

AI assistant

Search

Database backups