Tip: select any text on this page to bring up the Ask tooltip and question the Synedrix OS assistant about it.
What Grafana LGTM Stack is
The Grafana LGTM Stack is not one product. It is five separately released open-source projects from Grafana Labs designed to compose into a single observability backend: Loki for logs, Grafana for visualisation, Tempo for traces, and Mimir for metrics, with Alloy as the collector that feeds them.
Grafana's documentation uses a four-pillar framing rather than the more commonly cited three: metrics, logs, traces, and profiles. Synedrix OS adopts three of the four; profiling is not a canonical platform.
OpenTelemetry is the common ingestion standard across all of them. Alloy is an OpenTelemetry Collector distribution; Mimir ingests Prometheus or OpenTelemetry metrics; Loki natively supports OTLP over HTTP; Tempo accepts OTLP over gRPC and HTTP.
What it is normally used for
How Grafana LGTM Stack is used generally, before Synedrix OS enters the picture.
- Infrastructure and service dashboards over time-series data from many sources
- Log aggregation and search across distributed services
- Distributed tracing, with pivots from a slow trace into the logs around it
- Long-term, horizontally scalable metrics storage backed by object storage
- Alerting and SLO tracking with routing to on-call destinations
What it owns, and what it does not
Every platform in the stack has a non-overlapping responsibility. The boundary is as much a part of the design as the capability.
Owns
- Infrastructure and service health for every platform in the stack
- Dashboard rendering for the canonical metric families
- Threshold breaches and alert routing
- SLOs and their default thresholds
- Correlation carriage in alerts, so responders can pivot into the underlying trace
Does not own
- Is the model layer behaving? Traces, prompts, evaluations, cost
- → Langfuse (Layer 10)
- Is the product working for users? Behaviour, flags, experiments
- → PostHog (Layer 10)
- Raw high-volume time-series truth
- → QuestDB (Layer 7)
- Under whose authority did this happen?
- → Neo4j (Layer 7): lineage is a graph, not a trace
- The audit-grade record
- → Supabase PG17 (Layer 7)
Why Grafana LGTM Stack
It answers "is the system healthy?" (a different question from the other two)
Having three separate observability tools only makes sense if each one answers something genuinely different. This one answers whether the infrastructure itself is healthy; how the AI models are behaving and how the product is being used are answered by the other two.
Not locked into any one vendor's format
Every part of this stack accepts data in the OpenTelemetry format, an open standard. That means the code that produces telemetry doesn't need to change if a backend is ever swapped out.
One shared storage system for logs, traces, and metrics
Loki, Tempo, and Mimir (which handle logs, traces, and metrics respectively) all store their data in ordinary S3-compatible object storage. That means all three can share one local storage system in a self-hosted or air-gapped deployment, instead of needing three separate ones.
Metrics, logs, and traces are all one click apart
Metrics, logs, and traces each have their own query language, but all three share the same labelling and are all reachable from one Grafana screen. Moving from a metric to its logs to its trace is a click, not a switch to a different tool.
Licence position
Grafana LGTM Stack questions, answered
Short, citable answers drawn from this page.
What role does Grafana LGTM Stack play in Synedrix OS?
Grafana LGTM Stack (Operational monitoring) is mapped to Layer 10: Observability in Synedrix OS. Is the system healthy? Five projects composing into one observability backend.
What does Grafana LGTM Stack own in Synedrix OS?
Grafana LGTM Stack owns: Infrastructure and service health for every platform in the stack; Dashboard rendering for the canonical metric families; Threshold breaches and alert routing; SLOs and their default thresholds; Correlation carriage in alerts, so responders can pivot into the underlying trace.
What does Grafana LGTM Stack not own in Synedrix OS?
Grafana LGTM Stack does not own: Is the model layer behaving? Traces, prompts, evaluations, cost (owned by Langfuse (Layer 10)); Is the product working for users? Behaviour, flags, experiments (owned by PostHog (Layer 10)); Raw high-volume time-series truth (owned by QuestDB (Layer 7)); Under whose authority did this happen? (owned by Neo4j (Layer 7): lineage is a graph, not a trace); The audit-grade record (owned by Supabase PG17 (Layer 7)).
What licence is Grafana LGTM Stack under?
Grafana LGTM Stack is released under Grafana, Loki, Tempo, Mimir AGPL-3.0; Alloy Apache-2.0. Licences were verified individually and are not uniform. Four of the five components are AGPL-3.0, whose network-use clause is material for a commercial product: running unmodified upstream images is the low-friction path, while forking or patching triggers a source-offer obligation.
The full reference page, with the operating model, deployment detail, and links to every canonical specification this platform touches, lives in the documentation repository, which is private while the platform is being built.