Stability

Availability, recovery posture, SLOs and live incident control.

12 / 12 recovery tests passed

Availability

99.96%

rolling 30 days

Open incidents

1 high priority

MTTR

42 min

last 30 days

Recovery tests

12 / 12

passed this quarter

Target and actual performance for the core estate.

Objective	Target	Actual	State
Portal APIs	99.95%	99.99%	meeting
Model runtime	99.90%	99.94%	meeting
Evidence store	99.90%	99.97%	meeting
Training queue	95%	92%	watch
Decision API p95	250 ms	312 ms	breaching

Current stability signals with operating context attached.

High GPU queue

training pool eu-west-01

acknowledged

Model runtime patch

186 services compliant

complete

API latency watch

Decision API p95 high

investigate

Storage forecast

expand archive tier

planned