Enablement® — Service Availability

Availability Report · Advent Business Company Inc. · FedRAMP ID FR2628647239 · package id ADVENTBUSINESS-ENB · CDS-CSO-AVR · all times UTC · generated 2026-08-07T04:28:49Z · machine-readable JSON · certification package

Current status

Most recent probe of every measured endpoint succeeded. Last probe 2026-08-07T04:25:01Z (3.8 minutes ago). State code operational.

Last 30 days (2026-07-08T04:28:49Z → 2026-08-07T04:28:49Z)

uptimePercentOfObservedProbes = successful probes / probes actually taken. uptimePercentOfExpectedProbes = successful probes / probes that should have been taken over the measured span, i.e. a LOWER BOUND that treats an unrecorded probe as not-known-good. Read the lower bound as the conservative figure.

Availability incidents and measurement gaps

An availability incident is 2 or more consecutive failed probes on an endpoint (the shortest outage a 5-minute cadence can resolve), merged across endpoints when the failure windows overlap. A single isolated failed probe appears in the daily rollup but is not promoted to an incident.

Daily availability

What this measures

What it does not measure

ml_ai/service/security_scan/availability_probe.sh, host cron, every 5 minutes, writing to BPM_RUN.AVAILABILITY_PROBE with a 90-day retention sweep.

If this page will not load — the off-CSO mirror

Endpoint	Last probe	Result	HTTP
Public edge `public`	2026-08-07T04:25:01Z	reachable	200
Application tier `internal`	2026-08-07T04:25:01Z	reachable	200

Uptime (of probes taken)	94.585%
Uptime (lower bound, of probes expected)	94.166%
Probes	12139 successful of 12834 taken
History available	22.38 days (first probe 2026-07-15T19:20:02Z) — short of the 30-day requirement, filling by observation
Measurement coverage	99.558% of expected probes were actually recorded within that span (12891 expected at the 5-minute cadence)
Coverage of the full 30-day window	74.271% of the 17280 probes a complete window would contain. fullWindowCoveragePercent answers 'does the history reach back the full 30 days the rule requires'. measuredSpanCoveragePercent answers 'within the period the collector has been running, did it actually take every probe it should have'. A shortfall in the second is missing measurement, published as measurement-gap incidents, and is never counted as uptime.
Availability incidents	3
Measurement gaps	4

Endpoint	What it probes	Probes	Successful	Uptime	Median latency	Last probe
Public edge `public`	HTTPS front door at https://enablement.cc/ — the path every customer and agency user takes.	6417	6066	94.53%	31 ms	2026-08-07T04:25:01Z
Application tier `internal`	Tomcat probed directly on the host, bypassing the edge — separates an edge failure from an app failure.	6417	6073	94.639%	1 ms	2026-08-07T04:25:01Z

Type	From	To	Duration	Affected	Detail
Outage availability-incident	2026-07-30T03:15:02Z	2026-07-30T03:20:02Z	5.0 min First to last failed probe. The real outage began up to 5 minutes earlier and ended up to 5 minutes later than shown; this is the measured floor, not a precise duration.	Application tier (internal, 2 failed probes); Public edge (public, 2 failed probes) (full-service)	HTTP 502 automated 5-minute probe (no human report required)
Outage availability-incident	2026-07-20T03:25:02Z	2026-07-21T02:05:16Z	1360.2 min First to last failed probe. The real outage began up to 5 minutes earlier and ended up to 5 minutes later than shown; this is the measured floor, not a precise duration.	Application tier (internal, 272 failed probes); Public edge (public, 273 failed probes) (full-service)	HTTP 502 automated 5-minute probe (no human report required)
Outage availability-incident	2026-07-17T03:05:02Z	2026-07-17T03:10:16Z	5.2 min First to last failed probe. The real outage began up to 5 minutes earlier and ended up to 5 minutes later than shown; this is the measured floor, not a precise duration.	Public edge (public, 2 failed probes) (edge-only)	HTTP 502 automated 5-minute probe (no human report required)
Measurement gap measurement-gap	2026-07-16T14:20:01Z	2026-07-16T15:35:02Z	75.0 min	Public edge `public`	availability unknown No probe result exists for this period, so availability during it is unknown. It is NOT counted as uptime. A gap means the collector, the host, or the database was unavailable to record a result.
Measurement gap measurement-gap	2026-07-16T14:20:01Z	2026-07-16T15:35:02Z	75.0 min	Application tier `internal`	availability unknown No probe result exists for this period, so availability during it is unknown. It is NOT counted as uptime. A gap means the collector, the host, or the database was unavailable to record a result.
Measurement gap measurement-gap	2026-07-16T03:25:01Z	2026-07-16T04:40:02Z	75.0 min	Public edge `public`	availability unknown No probe result exists for this period, so availability during it is unknown. It is NOT counted as uptime. A gap means the collector, the host, or the database was unavailable to record a result.
Measurement gap measurement-gap	2026-07-16T03:25:01Z	2026-07-16T04:40:02Z	75.0 min	Application tier `internal`	availability unknown No probe result exists for this period, so availability during it is unknown. It is NOT counted as uptime. A gap means the collector, the host, or the database was unavailable to record a result.

Date (UTC)	Uptime	Probes OK / taken	Per endpoint	Coverage	Uptime lower bound
2026-08-07	98.113%	104 / 106	internal 52/53; public 52/53	100.0% of 106 expected	98.113%
2026-08-06	98.611%	568 / 576	internal 284/288; public 284/288	100.0% of 576 expected	98.611%
2026-08-05	99.306%	572 / 576	internal 286/288; public 286/288	100.0% of 576 expected	99.306%
2026-08-04	100.0%	576 / 576	internal 288/288; public 288/288	100.0% of 576 expected	100.0%
2026-08-03	99.132%	571 / 576	internal 286/288; public 285/288	100.0% of 576 expected	99.132%
2026-08-02	98.958%	570 / 576	internal 285/288; public 285/288	100.0% of 576 expected	98.958%
2026-08-01	100.0%	576 / 576	internal 288/288; public 288/288	100.0% of 576 expected	100.0%
2026-07-31	98.611%	568 / 576	internal 284/288; public 284/288	100.0% of 576 expected	98.611%
2026-07-30	97.569%	562 / 576	internal 281/288; public 281/288	100.0% of 576 expected	97.569%
2026-07-29	98.09%	565 / 576	internal 283/288; public 282/288	100.0% of 576 expected	98.09%
2026-07-28	96.875%	558 / 576	internal 279/288; public 279/288	100.0% of 576 expected	96.875%
2026-07-27	99.653%	574 / 576	internal 287/288; public 287/288	100.0% of 576 expected	99.653%
2026-07-26	99.479%	573 / 576	internal 287/288; public 286/288	100.0% of 576 expected	99.479%
2026-07-25	99.306%	572 / 576	internal 286/288; public 286/288	100.0% of 576 expected	99.306%
2026-07-24	100.0%	576 / 576	internal 288/288; public 288/288	100.0% of 576 expected	100.0%
2026-07-23	96.875%	558 / 576	internal 279/288; public 279/288	100.0% of 576 expected	96.875%
2026-07-22	99.306%	572 / 576	internal 286/288; public 286/288	100.0% of 576 expected	99.306%
2026-07-21	90.451%	521 / 576	internal 261/288; public 260/288	100.0% of 576 expected	90.451%
2026-07-20	14.236%	82 / 576	internal 41/288; public 41/288	100.0% of 576 expected	14.236%
2026-07-19	100.0%	576 / 576	internal 288/288; public 288/288	100.0% of 576 expected	100.0%
2026-07-18	97.917%	564 / 576	internal 282/288; public 282/288	100.0% of 576 expected	97.917%
2026-07-17	97.917%	564 / 576	internal 283/288; public 281/288	100.0% of 576 expected	97.917%
2026-07-16	97.885%	509 / 520	internal 255/260; public 254/260	90.278% of 576 expected	88.368%
2026-07-15	96.429%	108 / 112	internal 54/56; public 54/56	100.0% of 110 expected	98.182%
2026-07-14	not measured — availability unknown
2026-07-13	not measured — availability unknown
2026-07-12	not measured — availability unknown
2026-07-11	not measured — availability unknown
2026-07-10	not measured — availability unknown
2026-07-09	not measured — availability unknown
2026-07-08	not measured — availability unknown

https://d3arb1q8mam5wu.cloudfront.net
Bookmark it now, because during an outage you cannot reach this page to find it. Live external observation: https://d3arb1q8mam5wu.cloudfront.net/mirror-availability.json · mirrored copy of this document: https://d3arb1q8mam5wu.cloudfront.net/availability.json

A point-in-time copy of every PUBLIC trust-center artifact, refreshed every 15 minutes by a scheduled job running OUTSIDE this offering, in the AWS commercial partition (account 971328156383, us-east-1), served from S3 through CloudFront. It shares no partition, account, region, host, database, web server, TLS certificate or DNS zone with production.

Two different things, and the difference matters. (1) The mirrored copy of THIS document is last-known-good: it is stamped with the time it was taken and will not contain the outage in progress. (2) mirror-availability.json is LIVE during the outage — the refresher is itself an external probe, so when it cannot reach this offering it publishes that fact, with a first-failure time and a running duration, from infrastructure the outage does not touch. That second document is the one this on-host service structurally cannot produce.

https://d3arb1q8mam5wu.cloudfront.net/mirror-manifest.json carries a mirroredAt timestamp and a SHA-256 per artifact, and each stored object repeats it in an x-amz-meta-mirrored-at header. An age far past 15 minutes while the mirror reports this offering reachable means the refresher itself has stopped; that condition is alarmed by the same freshness worker that watches the rest of the evidence pipeline, so a silently frozen mirror is not left to be noticed by a reader. Only artifacts that are already public without authentication. No token-gated certification data is mirrored — not the evidence index, the package documents, nor the SDR, KSI, VDR, AVI, OCR, historical, assessor or query services. A static bucket has no token gate, so the copy list is an explicit allow-list of nine object keys and the storage policy grants read on exactly those object ARNs with no wildcard.

Limitations — read this

This report cannot report an outage it is part of.
This is a self-hosted availability report. The probe collector, the database it writes to, and the service that renders this document all run on the same production host in AWS GovCloud us-gov-east-1 as the offering being measured. It is not an independent, off-CSO status service.

Because this endpoint is served by the same application tier it reports on, it is unreachable during precisely the outages it exists to report. Treat a failure to reach this endpoint as itself an availability signal — the most severe one it can give you — and then read the off-CSO mirror at https://d3arb1q8mam5wu.cloudfront.net/mirror-availability.json, which is written from outside this infrastructure and stays up when this endpoint does not.

2026-07-20 production outage — the known instance.
A ~23-hour HTTP 502 took the entire enablement.cc origin down, trust center included. The probe recorded it: 247 consecutive failed probes per endpoint on 2026-07-20 and a further 27-28 on 2026-07-21. The data is therefore honest about the outage — but had this endpoint existed at the time, it could not have served that fact while it was true, because it was down with everything else. That is the limitation above, demonstrated rather than asserted.

The event is separately recorded in Advent's incident record and is not knowable only from this endpoint.

CDS-CSO-AVR compliance, arm by arm

Missing measurement is not uptime	A failed probe is positive evidence of an outage. An absent probe is NOT evidence of uptime. Missing measurement is published as measurement-gap incidents with availability 'unknown' and is excluded from the uptime numerator and denominator alike.
Not independently attested	Provider-generated. No third party attests these figures. An agency requiring independent availability evidence should operate its own external probe against https://enablement.cc/ and compare.

Requirement	Status	Basis
maintain a web service indicating current and historical availability	MET `met`	This endpoint plus the human-readable page at https://enablement.cc/ml/20x/status.
core services, over at least the past 30 days	PARTIAL `partial`	Continuous measurement began 2026-07-15; 22.4 days of history are available against a 30-day floor. The window fills by observation, not by backfill.
including availability incidents	MET `met`	Incidents are derived from the probe series and published in the incidents array; measurement gaps are published separately as unknown periods.
in both human-readable and machine-readable formats	MET `met`	JSON at https://enablement.cc/ml/20x/availability and HTML at https://enablement.cc/ml/20x/status, both rendered from this same document.
available to all necessary parties	MET `met`	Both formats are public and unauthenticated; no token is required.
MUST be available even if the primary cloud service offering is unavailable	MET `met`	Satisfied by the off-CSO mirror at https://d3arb1q8mam5wu.cloudfront.net — see limitations.offCsoMirror. THIS endpoint still shares a host, a database and an origin with the offering and still disappears with it; that has not changed and is not being claimed otherwise. What closes the arm is that the report no longer exists only here: a copy is refreshed every 15 minutes into the AWS commercial partition (account 971328156383, us-east-1) and served from S3 via CloudFront, sharing no partition, account, region, host, database, web server, TLS certificate or DNS zone with production. Verified by loading every mirrored artifact in a browser with all network access to enablement.cc refused. The qualification, stated rather than buried: the mirrored copy of this document is last-known-good and will lag an outage in progress, while the companion mirror-availability.json is live during one, because the refresher is an external probe and publishes its own failure to reach this offering.

Overall: PARTIAL (partial). Every content, format, access and independent-hosting arm of CDS-CSO-AVR is satisfied. The 30-day window is the only arm still filling: 22.4 days of continuous history exist against a 30-day floor, and the remaining 7.6 days close by observation, not by backfill. No wording here should be read as claiming otherwise.

Enablement® Service Availability