Strata · open source, self-hosted
Strata unites MonoFS, Guardian, Doctor, and Agent — open-source engines for your source, deployments, telemetry, and assistance. Nightwatch watches alerts in the background, investigates them, and hands you the evidence. Apache-2.0, self-hosted, and developed in the open.
Sentinel alerting + Nightwatch. Alerts now arrive with evidence, confidence, and a handoff instead of a wall of noise.
See NightwatchThe platform
Most platforms are a pile of tools with a shared login. Strata is built on a shared substrate — the same files, state, and telemetry underneath every layer. The whole stack is open source and runs on your own hardware.
Source, artifacts, deployment state, and telemetry share one durable substrate. No copy jobs, no sync drift.
Every trace and metric can be read next to the rollout that produced it. “What changed?” stops being a hunt.
FUSE, OTLP, OCI, Prometheus, Loki, and Zipkin. Strata connects the standards you already run.
Desired state is diffed and checked before it is applied. Drift is visible, mutations are attributable.
Sentinel alerts, deployment events, traces, and topology meet in one incident view, so the first answer is evidence, not a guess.
The stack
Each engine owns a hard problem. The magic is what happens when they share the same fabric.
One workspace. Every repo.
Mount thousands of independent repositories as one writable virtual monorepo. Edit locally, publish explicitly, and keep the whole platform's bytes in one place.
monofs-session commitDeclare it. Guardian lands it.
An intent-based control plane that compiles desired state into a dependency graph, detects drift, and applies the delta across Kubernetes, Docker, AWS, or local targets.
Every signal. One story.
A privacy-first OpenTelemetry backend that stores traces, logs, and metrics together — then adds Guardian deployment context and Grafana-compatible APIs.
Ask the platform. Approve the change.
An assistant with live Guardian, Doctor, and MonoFS tools. It answers from evidence instead of guessing, plans the next change, and keeps every mutation behind an approval gate.
Nightwatch
Turn on one partition label. Nightwatch watches alerts, investigates with Doctor, and hands you the incident with the evidence already attached. Shadow mode by default — it never gets the keys.
Per-partition opt-in.
A single partition label turns Nightwatch on. It polls Sentinel alerts, groups related signals into incidents, and keeps a shift log of what happened while you were away.
Metrics, traces, topology, events.
Nightwatch builds an evidence pack from Doctor and Guardian: recent deploys, service topology, trace edges, log patterns, and fleet metric anomalies. The LLM explains the evidence; it does not invent it.
Read globally. Act locally.
If the root cause is outside the assigned partition, Nightwatch creates a cross-partition finding and hands it to the owning team. Low confidence, missing evidence, or a failed investigation becomes an escalation recommendation.
How it flows
One loop. No hand-offs between tools. This is the path a change takes through Strata.
A developer opens a virtual monorepo, edits in a local overlay, and publishes one explicit commit. No giant clone, no waiting for an indexer.
MonoFSGuardian compiles the dependency graph, shows the drift, checks that the change is safe, then applies it in the right order.
GuardianTraces, logs, and metrics land in Doctor alongside the deployment event that produced them. The rollout is part of the trace.
DoctorAgent answers from live Guardian, Doctor, and MonoFS evidence, plans the next change, and keeps mutations behind approval.
AgentAlerts arrive, Nightwatch correlates them with deploys and telemetry, and hands over a readable incident with evidence, confidence, and an escalation when it cannot be sure.
NightwatchThe primitives
Strata is a surface, not a proprietary layer. Underneath are filesystems, manifests, and standard protocols you can inspect, script, and take with you.
The snippets are the actual interfaces. Swap the hostnames for yours.
monofs-client --mount=/dev/monorepo \
--router=monofs:9090 \
--virtual-monorepo --writable
# work normally — git, IDEs, tests
cd /dev/monorepo && git status
# publish explicitly
monofs-session diff
monofs-session commit -m "feat: ship the dashboard"apiVersion: guardian/v1alpha1
kind: Partition
metadata:
name: production
spec:
intents:
- name: database
targetPusher: k8s-prod
- name: api-server
targetPusher: k8s-prod
joins: [database]
# drift is computed before anything is applied
guardianctl partition reconcile --partition productionexport OTEL_EXPORTER_OTLP_ENDPOINT=http://doctor:4317
export OTEL_SERVICE_NAME=api-server
# traces, logs, and metrics share one store
curl "http://doctor:18080/v1/traces/<traceID>"
# read rollout health next to the telemetry
curl "http://doctor:18080/v1/guardian/health?partition=production"cd agent && docker compose up -d
# UI :3000 · API :8090
curl -s http://localhost:8090/api/chat \
-H "content-type: application/json" \
-d '{"message":"why did production drift?"}'
# answers cite live state; mutations stay behind approvalWhy Strata
MonoFS is the substrate under Guardian, Doctor, and Agent. That means fewer copies, fewer credentials, and one place to reason about durability.
Doctor ingests Guardian events directly. Nightwatch correlates alerts with traces, topology, and recent changes before it hands over an incident.
Drift is detected, changes are validated, and every write carries a principal, reason, and correlation ID.
Doctor anonymizes sensitive fields with deterministic HMAC tokenization, per tenant, before persistence.
OTLP collectors, Grafana dashboards, OCI images, S3 or GCS upstreams. Strata fits the stack you have.
Kubernetes, Docker, AWS, or a single local process. Same intent model, same telemetry surface, same audit trail — wherever the work lands.
Open source
No sales call and no waiting list. Clone the repos, run the stack, and open an issue when something breaks. Issues and PRs are welcome.
github.com/radryc/strata · Apache-2.0 · open source