Mastering Log Cost Control with Adaptive Logs Drop Rules: A Q&A Guide

Platform and observability teams often grapple with noisy logs—health check messages, forgotten debug statements, or verbose info logs that inflate costs without adding value. Until recently, removing these logs required cumbersome infrastructure changes. With the public preview of drop rules in Adaptive Logs, you can now define custom rules to drop low-value logs before they’re written to Grafana Cloud Logs, instantly reducing noise and saving money. This Q&A covers how drop rules work, what you can achieve, and how they fit into a complete log cost management strategy.

1. What are Adaptive Logs drop rules and why are they important?

Drop rules let you create custom logic—using log labels, detected log levels, or line content—to discard logs before ingestion. This is crucial because most teams have logs they know are noise but lack an easy way to eliminate them. Centralized teams want control without toilsome change management. Drop rules provide a simple, rule-based mechanism to drop unwanted logs right at the edge of Grafana Cloud, reducing storage and query costs. This capability, already available in Adaptive Metrics and Adaptive Traces, now extends to logs, giving you a unified toolset for optimization.

2. How do drop rules help reduce noise and costs?

By dropping logs before they’re stored, you avoid paying for ingestion, retention, and querying of low-value data. For example, a single rule can drop 100% of health check logs across all services, ensuring only meaningful logs are retained. Drop rules are evaluated in priority order—once a rule matches, its drop rate applies. This eliminates noisy logs instantly, freeing budget for high-value observability data. The result: a cleaner, more cost-effective log environment without requiring developers to change application logging.

3. What kinds of logs can be dropped using drop rules?

Drop rules offer flexible criteria. You can target logs by:

• Log level – e.g., drop all DEBUG logs that often flood your budget.
• Log content – e.g., drop lines containing a specific string like “healthcheck” or “heartbeat”.
• Labels – e.g., target a particular service or environment with a label selector.
• Combinations – e.g., drop DEBUG logs from a specific service that also match a text pattern.

This granularity lets you surgically remove noise while preserving important logs.

4. Can I sample logs instead of dropping them entirely?

Yes. Drop rules allow you to specify a drop percentage, effectively sampling noisy logs you don’t want to discard completely. For instance, you can apply a 90% drop rate to a batch processing job that generates repetitive logs. This keeps a representative 10% for troubleshooting while cutting costs by 90%. Sampling is ideal when you need some visibility but can tolerate reduced fidelity.

5. How do drop rules interact with exemptions and patterns?

Adaptive Logs processes log lines in three stages:

1. Exemptions – Logs that match an exemption pass through untouched, ignoring sampling.
2. Drop rules – Evaluated in priority order; the first matching rule applies its drop rate.
3. Patterns – Optimization recommendations (adaptive patterns) apply to logs not exempted or dropped.

This layered approach lets you protect critical logs, drop known noise, and then intelligently sample remaining volume. See our use cases to see how they work together.

6. What are some practical use cases for drop rules?

• Drop by level: Eliminate all DEBUG logs across every service with one rule.
• Sample chatty logs: Apply an 80% drop rate to repetitive health check logs from a Kubernetes cluster.
• Target noisy producers: A legacy service suddenly emits high-volume INFO logs. Use a label selector plus content filter to drop 90% of them.
• Environment-specific: Drop verbose logs from staging but keep them in production.

Each rule can use multiple criteria for precision.

7. How do I get started with drop rules in Grafana Cloud?

Navigate to Adaptive Logs in your Grafana Cloud account. Under the “Drop Rules” tab, click “Add rule”. Define your criteria—log level, label selectors, text patterns—and set a drop percentage (0-100%). Save the rule; it takes effect immediately for new logs. Review the official documentation for detailed examples and best practices. Start with a small drop rate to test, then adjust as needed.