Update dependencies to Go 1.24, add support for snapshot envelopes, a…

.gitattributes

@@ -0,0 +1,2 @@
+api/schema/*.json text eol=lf
+internal/schema/schema/*.json text eol=lf

README.md

@@ -1,45 +1,70 @@
 # Bering
 
-Bering builds a resilience model artifact from trace artifacts. The output artifact is `bering-model.json` in `BeringResilienceModel` v1.0.0 format. Use it in any downstream tooling or analytics pipeline.
+Bering is a discovery and publishing layer for service topology and endpoint contracts.
 
-## Contract
+It supports two operating modes:
 
-Bering pins `metadata.schema` exactly:
+- deterministic batch discovery from trace files and directories
+- long-running runtime discovery that accepts OTLP/HTTP spans and publishes rolling snapshot envelopes for observability consumers
+
+Bering owns discovery and discovery-side public contracts. It does not own simulation, gating, chaos execution, or policy decisions.
+
+## Public Artifacts
+
+Bering currently publishes two versioned JSON artifact types.
+
+### Core model
 
 - `name`: `io.mb3r.bering.model`
 - `version`: `1.0.0`
 - `uri`: `https://mb3r-lab.github.io/Bering/schema/model/v1.0.0/model.schema.json`
 - `digest`: `sha256:272277c093f37580adcd2dded225bd37c86539d642d7910baad7e4228227d1a7`
 
-Any mismatch fails validation.
+This remains the simple stable topology artifact for file-based users and downstream tools such as Sheaft.
+
+### Snapshot envelope
+
+- `name`: `io.mb3r.bering.snapshot`
+- `version`: `1.0.0`
+- `uri`: `https://mb3r-lab.github.io/Bering/schema/snapshot/v1.0.0/snapshot.schema.json`
+- `digest`: `sha256:87e4e887ed4a37b72f6136e268b73552eccb92941c4de2c6f3a514dd066ea972`
+
+This wraps the core model with runtime window metadata, ingest counts, support summaries, provenance, and topology diffs.
+
+`bering validate` accepts either artifact type.
 
 ## Repository layout
 
 ```text
 cmd/bering                    CLI entrypoint
 internal/app                  command wiring
-internal/connectors/traces    trace file/dir loading and normalization
-internal/discovery            model inference from normalized spans
-internal/model                model structs, semantic checks, canonical IO
+internal/config               serve-mode config parsing and validation
+internal/connectors/traces    file/dir trace loading and normalization
+internal/connectors/otlp      OTLP/HTTP request decoding into normalized spans
+internal/discovery            source-agnostic discovery engine and overlay application
+internal/model                stable core model structs, semantic checks, canonical IO
+internal/overlay              generic discovery overlay loader
+internal/runtime              long-running service, tumbling windows, sinks, metrics
 internal/schema               pinned contract constants + JSON Schema validation
-internal/jsoncanon            deterministic recursive JSON encoder
-api/schema                    versioned public schema
-configs                       sample configs (replicas override)
-examples                      trace fixtures + expected output artifacts
-docs                          format, heuristic, and MVP limits
+internal/snapshot             snapshot envelope structs, diffing, canonical IO
+api/schema                    public schemas published via GitHub Pages
+configs                       sample serve and overlay configs
+examples                      traces, outputs, collector/prometheus/grafana examples
+docs                          architecture, contract, config, migration, limits
 scripts/ci                    CI helper scripts
 ```
 
 ## Commands
 
 ```bash
-bering discover --input <trace-file|dir> [--out bering-model.json] [--replicas replicas.yaml|json] [--discovered-at RFC3339]
-bering validate --input <bering-model.json>
+bering discover --input <trace-file|dir> [--out bering-model.json] [--snapshot-out bering-snapshot.json] [--replicas replicas.yaml|json] [--overlay overlay.yaml] [--discovered-at RFC3339]
+bering validate --input <bering-model.json|bering-snapshot.json>
+bering serve --config configs/serve.sample.yaml [--listen :4318] [--window-size 30s] [--flush-interval 5s]
 ```
 
 ## Quickstart
 
-### 1) Discover from traces
+### 1) Batch discovery from traces
 
 ```bash
 go run ./cmd/bering discover \
@@ -48,59 +73,119 @@ go run ./cmd/bering discover \
   --discovered-at 2026-03-03T00:00:00Z
 ```
 
-### 2) Validate artifact
+### 2) Validate the model artifact
 
 ```bash
 go run ./cmd/bering validate \
   --input examples/outputs/bering-model.normalized.sample.json
 ```
 
-### 3) Use the model in any downstream tool
+### 3) Generate a snapshot envelope in batch mode
 
-Examples:
+```bash
+go run ./cmd/bering discover \
+  --input examples/traces/normalized.sample.json \
+  --out out/bering-model.json \
+  --snapshot-out examples/outputs/bering-snapshot.normalized.sample.json \
+  --overlay configs/discovery.overlay.sample.yaml \
+  --discovered-at 2026-03-03T00:00:00Z
+```
 
-- run your own analytics (topology checks, risk scoring, SLO diagnostics)
-- feed the model into simulation tooling
-- simulate failures with [Sheaft](https://github.com/MB3R-Lab/Sheaft) (one possible consumer), for example:
-  ```bash
-  # from sibling Sheaft repository
-  go run ./cmd/sheaft run \
-    --model ../Bering/examples/outputs/bering-model.normalized.sample.json \
-    --policy configs/gate.policy.example.yaml \
-    --out-dir out \
-    --seed 42
-  ```
+### 4) Run the runtime service
 
-## Deterministic output
+```bash
+go run ./cmd/bering serve --config configs/serve.sample.yaml
+```
 
-Bering output is deterministic for identical inputs and flags:
+The runtime service exposes:
+
+- `POST /v1/traces` for OTLP/HTTP trace ingest
+- `GET /healthz`
+- `GET /readyz`
+- `GET /metrics`
+
+The primary integration path is standard OpenTelemetry Collector or SDK exporters sending spans to Bering over OTLP/HTTP. No custom Collector build is required.
+
+### 5) Use the stable model with Sheaft
+
+```bash
+# from a sibling Sheaft repository
+go run ./cmd/sheaft run \
+  --model ../Bering/examples/outputs/bering-model.normalized.sample.json \
+  --policy configs/gate.policy.example.yaml \
+  --out-dir out \
+  --seed 42
+```
+
+## Examples
+
+- Batch inputs: [examples/traces/normalized.sample.json](examples/traces/normalized.sample.json), [examples/traces/otel.sample.json](examples/traces/otel.sample.json)
+- Batch outputs: [examples/outputs/bering-model.normalized.sample.json](examples/outputs/bering-model.normalized.sample.json), [examples/outputs/bering-snapshot.normalized.sample.json](examples/outputs/bering-snapshot.normalized.sample.json)
+- Runtime config: [configs/serve.sample.yaml](configs/serve.sample.yaml)
+- Discovery overlay: [configs/discovery.overlay.sample.yaml](configs/discovery.overlay.sample.yaml)
+- Collector sidecar: [examples/collector/otelcol.sidecar.yaml](examples/collector/otelcol.sidecar.yaml)
+- Prometheus scrape config: [examples/prometheus/bering.prometheus.yml](examples/prometheus/bering.prometheus.yml)
+- Grafana dashboard: [examples/grafana/bering-runtime-dashboard.json](examples/grafana/bering-runtime-dashboard.json)
+
+## Determinism and Runtime Tradeoffs
+
+Batch output remains deterministic for identical inputs and flags:
 
 - services sorted by `id`
 - edges sorted by `(from,to,kind,blocking)`
 - endpoints sorted by `id`
-- stable IDs for services/edges/endpoints
-- canonical JSON writer with recursive object-key ordering (future-safe for map fields)
+- canonical JSON output with stable object-key ordering
 - optional `--discovered-at` for reproducible timestamps
 
-## Supported trace formats (MVP)
+Runtime mode is intentionally bounded, not lossless:
+
+- one active tumbling window is retained in memory
+- the previous emitted snapshot is retained for diffs and carry-forward runtime timestamps
+- `runtime.max_in_memory_spans` bounds retained spans per active window
+- late spans follow `drop` or `current_window` policy
+- spans beyond the configured in-memory cap are dropped and surfaced via metrics/logs
+- empty windows are advanced without emitting empty snapshots
+
+## Discovery overlays
+
+Discovery overlays are additive metadata inputs with explicit precedence by file order. They are intended for discovery-side enrichment, not policy evaluation.
+
+Supported examples include:
+
+- service labels and failure-eligibility labels
+- endpoint predicate references
+- workload or endpoint weights
+- SLO references or tags
+- replica overrides
 
-- Normalized JSON: `{"spans": [...]}` payload with canonical span fields.
-- Raw OTel JSON: `resourceSpans/scopeSpans/spans` payload.
+See [configs/discovery.overlay.sample.yaml](configs/discovery.overlay.sample.yaml).
 
-Details: [docs/trace-input-format.md](docs/trace-input-format.md)
+## Metrics
 
-## Schema publishing
+The runtime service exports Prometheus/OpenMetrics metrics including:
 
-Schema publishing is automated via GitHub Pages and release tags.
+- `spans_ingested_total`
+- `spans_dropped_total`
+- `snapshots_emitted_total`
+- `snapshot_build_duration_seconds`
+- `current_services`
+- `current_edges`
+- `current_endpoints`
+- `window_lag_seconds`
+- `last_snapshot_unixtime`
+- `snapshot_age_seconds`
+- `diff_added_*`
+- `diff_removed_*`
+- `diff_changed_*`
 
-- Workflow: `.github/workflows/publish-schema.yml`
-- Trigger: tags matching `schema-v*` (for example `schema-v1.0.0`)
-- Published paths:
-  - `https://mb3r-lab.github.io/Bering/schema/model/v1.0.0/model.schema.json`
-  - `https://mb3r-lab.github.io/Bering/schema/model/latest/model.schema.json`
-  - `https://mb3r-lab.github.io/Bering/schema/index.json`
+## Additional docs
 
-Operational steps are documented in [docs/schema-publishing.md](docs/schema-publishing.md).
+- [docs/architecture.md](docs/architecture.md)
+- [docs/runtime-config.md](docs/runtime-config.md)
+- [docs/trace-input-format.md](docs/trace-input-format.md)
+- [docs/schema-publishing.md](docs/schema-publishing.md)
+- [docs/migration-notes.md](docs/migration-notes.md)
+- [docs/mvp-scope-and-limits.md](docs/mvp-scope-and-limits.md)
 
 ## CI and local checks