Logging Configuration¶

The Logging tab allows you to control the verbosity level of the agent's internal logging system. Choosing the right logging level is important to balance operational visibility with performance and storage efficiency — particularly in production environments where high log volumes can affect throughput.

To access this tab, open the Agent Configuration dialog and select Logging from the tab bar.

Log Levels¶

BizMetry supports the following logging levels, from most to least verbose. Each level includes all messages from the levels above it in severity — for example, WARN also captures ERROR and FATAL messages.

TRACE¶

The most granular logging level available. Captures every internal operation the agent performs, including low-level execution paths, internal state transitions, and detailed protocol-level events.

When to use: Only for deep troubleshooting scenarios where DEBUG alone does not provide enough context to diagnose a problem. Expect very high log volume — do not use in production. Always revert to a higher level once the investigation is complete.

DEBUG¶

Detailed diagnostic information useful for understanding the agent's internal behavior during development or troubleshooting. Includes information about metric pipeline operations, sync cycles, connection attempts, and internal decision logic.

When to use: When investigating a specific issue in a non-production environment and you need more context than INFO provides. Acceptable for short periods in staging or QA environments. Avoid in production unless strictly necessary and for the shortest possible time.

INFO¶

General operational messages that describe the normal lifecycle of the agent — startup, sync cycles, configuration reloads, and connectivity events. This is the recommended level for all environments under normal operating conditions.

When to use: Always, unless you have a specific reason to change it. INFO provides enough context to understand the agent's operational state without generating excessive log volume.

WARN¶

Captures warning messages that indicate something unexpected occurred, but the agent was able to continue operating normally. Examples include temporary connectivity issues, near-capacity buffer usage, or configuration values approaching their limits.

When to use: When you want to reduce log noise in a stable production environment while still retaining visibility into conditions that may require attention before they escalate into errors.

ERROR¶

Captures only error-level events — conditions that caused an operation to fail but did not stop the agent from running. Examples include failed sync attempts, metric transmission errors, or rejected configuration updates.

When to use: In mature production environments where the agent is stable and you only want to be alerted to actual failures. Combine with a log monitoring tool that triggers alerts on ERROR entries.

FATAL¶

The most restrictive error level. Only captures critical failures that cause the agent to terminate or enter an unrecoverable state.

When to use: Rarely chosen directly. Appropriate only when you are confident the environment is fully stable and you want to minimize logging to the absolute minimum while still capturing crash-level events.

OFF¶

Disables all logging entirely. No log output is produced by the agent.

When to use: Not recommended for any operational environment. May be appropriate in highly constrained resource scenarios where storage is extremely limited, but note that any issue that occurs while logging is off will produce no evidence for diagnosis.

ALL¶

Enables all logging levels simultaneously — equivalent to TRACE but making the intent explicit. Captures every possible log entry the agent can produce.

When to use: Only during controlled debugging sessions in isolated, non-production environments. Log volume will be very high. Always revert to INFO or higher once the session is complete.

Recommendations¶

Use INFO as your baseline Under normal operating conditions, INFO is the recommended level for all environments. It provides a clear picture of the agent's lifecycle without generating log pressure.

Never run TRACE or DEBUG in production long-term These levels generate high log volumes that increase disk I/O, consume Log Buffer space faster, and can degrade agent throughput. If you need to activate them in a production agent for diagnostic purposes, do so for the shortest possible window and revert immediately.

Use ERROR or WARN in stable production environments Once an agent has been running reliably for an extended period, raising the log level to WARN or ERROR reduces log pressure and frees up Log Buffer capacity for metric-intensive workloads.

Match the log level to the environment type As a general rule of thumb, align log levels with environment tiers:

Monitor Log Pressure Regardless of the level chosen, always keep an eye on the Log Pressure indicator in the Agent Monitoring view. A rising Log Pressure value means the agent is generating logs faster than it can flush them. If this occurs while running at INFO, consider raising the level to WARN or reviewing the Log Batch Size configuration.

Rotate log levels during incident response When an incident occurs in production, the recommended approach is to temporarily lower the log level to DEBUG for the affected agent, capture the diagnostic information needed, and then revert to INFO or higher once the investigation is complete. Avoid leaving the agent at a verbose level after the incident is resolved.

Avoid OFF in any monitored environment Disabling logs entirely means losing all observability into the agent's behavior. If storage is a concern, prefer raising the level to FATAL rather than disabling logging altogether.

Important — Restart Required After Changes¶

Related Topics¶

• General Configuration
• SLA Configuration
• Agent Monitoring & Statistics
• Restarting an Agent