01
Know what changed
Connect deployments, runtime events, logs, and alerts so incidents have context.
ObservabilityMetricsLogsTracesAlerts
See the system.Read the signal.Fix faster.
Forgeon Observability brings metrics, logs, traces, alerts, runtime health, deployment context, and incident history into one interactive view of your applications.
Signal map
click a service node to inspect
selected service
deploy-service
edge-sync: routed request to runtime deployment deploy-v42deploy-service: build queue pressure above normal thresholdruntime-service: container health check passed on port 3000worker: job retry scheduled after temporary upstream timeoutapigate: request completed status=200 duration=42msedge-sync: routed request to runtime deployment deploy-v42deploy-service: build queue pressure above normal thresholdruntime-service: container health check passed on port 3000worker: job retry scheduled after temporary upstream timeoutapigate: request completed status=200 duration=42msedge-sync: routed request to runtime deployment deploy-v42deploy-service: build queue pressure above normal thresholdruntime-service: container health check passed on port 3000worker: job retry scheduled after temporary upstream timeoutapigate: request completed status=200 duration=42ms
Signal layers
Logs tell you what happened. Traces tell you where. Metrics tell you how bad.
Observability is not one chart. It is the overlap between symptoms, events, request paths, deployment context, runtime health, and user impact.
runtime awaredeployment awareincident readytrace connected
02
Find the hot path
Use traces and latency spans to see where requests are actually spending time.
03
Catch silent failures
Surface unhealthy runtimes, error spikes, crash loops, and queue pressure before users shout.
Trace waterfall
request id req_7f41ac
edge.resolve_domain
edge-sync
8ms
gateway.authenticate
apigate
12ms
runtime.forward
runtime-service
38ms
database.query
postgres
61ms
response.stream
edge-sync
9ms
Request path
Stop guessing where the request got slow.
A trace shows the request path as a chain of spans, so you can see whether the problem is at the edge, gateway, runtime, database, or downstream dependency.
Span timing
Measure where time disappears.
Context attached
Tie traces back to deploys and services.
Live logs
The truth usually appears first in the stream.
Logs should not be a dark forest. Filter by service, level, deployment, runtime instance, request id, or event type, then jump from a log line into traces and metrics.
logs.stream
12:41:08
infoedge-sync routed request to runtime deployment deploy-v42
12:41:11
warndeploy-service build queue pressure above normal threshold
12:41:16
inforuntime-service container health check passed on port 3000
12:41:22
errorworker job retry scheduled after temporary upstream timeout
12:41:25
infoapigate request completed status=200 duration=42ms
Incident replay
Incidents become easier when the story is already assembled.
Click through incident cards and connect symptoms to logs, traces, deploys, and runtime state. No more archaeology at 2 AM with cold coffee and emotional damage.
selected incident · 09:42
Latency spike detected
API p95 crossed 500ms for 4 minutes.
Observability
Make your platform explain itself.
Connect runtime metrics, logs, traces, deployments, alerts, and incidents so every problem comes with a trail, not a guessing game.
trace → log → metric → incident