Log output

Dependably writes structured logs to stdout. Every event is one object on one line, ready for Elastic Stack, AWS CloudWatch Logs, Datadog, Grafana Loki, or any other line-oriented log aggregator. Nothing needs to be enabled — the only choice is the format.

There is no in-app log viewer. Logs are write-only from the application's side; retrieval and search belong to whatever aggregator your deployment already runs. What Dependably does keep and serve in-app is the Audit log, which is a record of organization events, not of application diagnostics.

Choose a format

Set LOG_FORMAT (see Configuration):

Value	Use it for
json (default)	Elastic Common Schema (ECS) JSON, one object per line — log-ingestion pipelines.
text	Human-readable console output — interactive docker logs tailing.

Text format looks like this, with exception stack traces on the lines immediately following:

[2026-06-15 12:00:00.000 INF] Dependably.Api.NuGetController: Package pushed: my-pkg 1.0.0

The JSON fields

With LOG_FORMAT=json, each event conforms to the Elastic Common Schema:

Field	Always present	Description
@timestamp	yes	Event timestamp, ISO-8601 with offset.
log.level	yes	Verbose, Debug, Information, Warning, Error, or Fatal.
message	yes	The rendered message.
ecs.version	yes	ECS schema version.
log.logger	when set	Full type name of the logger.
trace.id, span.id	when set	Trace and span IDs from the current activity.
labels	varies	Every remaining property. Sensitive values are already redacted to [REDACTED] before the formatter sees them.

One event, formatted here for readability — real output is a single line:

{
  "@timestamp": "2026-06-15T12:00:00.000+00:00",
  "log.level": "Information",
  "message": "Package pushed: my-pkg 1.0.0",
  "ecs.version": "9.0.0",
  "log": { "logger": "Dependably.Api.NuGetController" },
  "labels": { "OrgId": "default", "Ecosystem": "nuget", "PackageName": "my-pkg", "Version": "1.0.0" }
}

What each level means

Level	Used for
Verbose	Fine-grained diagnostics. Off in production by default.
Debug	Periodic heartbeats and operational ticks — a cleanup pass starting, a stats refresh finishing.
Information	Normal operational events: pushes, proxy hits, first boot.
Warning	Recoverable anomalies: upstream timeouts, checksum mismatches, skipped schema migrations.
Error	Failed requests and exceptions caught at the boundary.
Fatal	Unrecoverable startup failures.

Ship them to CloudWatch

1. Run the container with the awslogs log driver (or Firelens/FluentBit) pointed at a CloudWatch log group.
2. Leave LOG_FORMAT at its json default.
3. Query the ECS fields directly in CloudWatch Logs Insights:

fields `@timestamp`, `log.level`, message, `log.logger`
| filter `log.level` = "Error"
| sort `@timestamp` desc

Ship them over OTLP

Set OTEL_EXPORTER_OTLP_ENDPOINT and logs are forwarded to that collector in addition to the always-on stdout sink — stdout never stops. Air-gapped deployments leave the variable unset and consume stdout only.

Metrics have their own path: see the Grafana dashboard guide.