D3 Reliability · OPS · P1 deployment

Operations & substrate

Proof/substrate evidence surfaced
Substrate10 service areas

Evidence surfaced; missing artifacts stay marked pending

Deploybuild → ship → reload

One script, all P-services

DrillsReplay · Chaos · tracked

Operator drills visible; production witness gate remains

Platform spine

P01–P10 services

P-idServiceImplementationData pathUnit / Dockerfile
P01IdentityKeycloakliveinfra/data/keycloak/shipped
P02TimeChrony NTS / GNSSliveservices/p02-time/shipped
P03Data spinePostgres 16 + Citusliveinfra/data/pg/shipped
P04BusNATS JetStream 2.10liveinfra/data/nats/shipped
P05AuditBLAKE3 hash-chainliveservices/p05-audit/migrations/shipped
P06RulesCedar + signed YAMLliveservices/p06-rules/shipped
P07DocMinIO WORM + Apicurioliveinfra/data/minio/shipped
P08UINext.js 15 + React 19liveweb/app/shipped
P09ObservabilityOpenTelemetry → Grafanaliveobservability/pending
P10AnalyticsClickHouse 24liveinfra/data/clickhouse/shipped

Deployment artefacts

Build → ship → reload

Dockerfileson diskservices/p01-identity/Dockerfile
systemd unitson diskinfra/systemd/
Caddy reverse-proxyon diskinfra/caddy/Caddyfile
Deploy scripton diskinfra/deploy.sh
CI workflowon disk.github/workflows/ci.yml
D3 reliability CIon disk.github/workflows/d3-reliability-ci.yml
D4 substrate CIon disk.github/workflows/d4-substrate-buf.yml

Substrate containers · live

Docker-compose health

8 / 8 healthy
ContainerImageRoleUptimeHealth
axiom-pgpostgres:16-alpineData spine (P03)2 hHealthy
axiom-natsnats:2.10-alpineBus (P04)2 hHealthy
axiom-clickhouseclickhouse:24.8Analytics (P10)2 hHealthy
axiom-miniominio:latestWORM doc store (P07)2 hHealthy
axiom-keycloakkeycloak:25.0Identity (P01)2 hHealthy
axiom-openbaoopenbao:latestKMS · secrets2 hHealthy
axiom-apicurioapicurio-registrySchema registry (P07)2 hHealthy
axiom-valkeyvalkey:8-alpineCache · sessions2 hHealthy

Recent deploys · build → ship → reload

Last 6 deploys

all green
DeployServiceCommitValidatorsDurationState
DEP-26050813-007p11-dispatch1f9af8e89/892m 14sReleased
DEP-26050812-006p13-techff9a33c89/892m 02sReleased
DEP-26050812-005p14-sms28fec6e89/891m 58sReleased
DEP-26050811-004p12-crew8fe3bec89/892m 09sReleased
DEP-26050811-003web-shell17304a989/890m 47sReleased
DEP-26050810-002p05-audit04f328d89/891m 32sReleased

Live drills

Phase-1 acceptance drills

≤ 10 min

Replay drill

Junior auditor reproduces a full flight cycle from P05 alone in under 10 minutes. Acceptance criterion for Phase 1 close.

evidence on diskreports/audit-replay-drill.log

P02 · P04 · P05 · P10

Chaos drill

Controlled outage of P02 (time), P04 (bus), P05 (audit), P10 (analytics). System degrades gracefully; no audit chain breaks; no data loss.

evidence on diskreports/chaos-drill.log
Operations & substrate | AXIOM