Fair usage limits: a safer way to scale observability

For the past several years, Coralogix customers have used the platform to ingest, process, and analyze large volumes of observability data without the presence of artificial barriers or unexpected constraints. This flexibility has enabled teams to experiment freely, evolve their architectures, and scale smoothly alongside their systems.

As usage grows, platform behavior at scale becomes increasingly important. We have observed that without explicit boundaries, unbounded usage patterns can create unexpected challenges for your own operations. These situations were not typically the result of incorrect usage, but rather a lack of visibility. Without clear feedback loops, a single noisy environment can unintentionally choke the allocation for others, or one high-volume team can deplete the shared resource pool for the rest of the organization. This often results in sudden spikes in unit consumption that are only discovered after the impact has occurred.

To address this, we are introducing fair usage limits. While we are starting with metrics and infrastructure monitoring, this framework is being built to cover all aspects of the product in the   near future and  is designed to improve transparency, predictability, and platform reliability as your workloads scale.

Why we made this change

The goal of fair usage limits is not to restrict, but rather to improve clarity. We know that “limits” is a word no one wants to hear, but this change is designed to be an upgrade to predictability.

By defining explicit thresholds for high-impact behaviors and metrics such as ingestion volume, label cardinality, and query complexity, we provide teams with better insight into how the system behaves at scale. This allows usage patterns to be better  understood and adjusted early, before they affect stability or performance.

Rather than interrupting workflows, these limits are designed to surface context. When thresholds are approached or exceeded, Coralogix emits structured diagnostic logs that explain what occurred, why it mattered, and what actions are available. This information is available directly within your observability data, making it possible to monitor and respond using the same tools you already rely on.

The intent is to support confident scaling through visibility and feedback, not enforcement for its own sake.

What is covered by fair usage limits

Fair usage limits apply to a set of platform behaviors that have a disproportionate impact on system performance and reliability.

These include:

• Ingestion limits, which help prevent runaway cardinality or excessive sampling from destabilizing ingestion pipelines
  
• Structural limits such as Metric Label Length, which ensure that excessively large or numerous labels do not result in silent drops or degraded performance
  
• Query limits, which protect dashboard and alert responsiveness by applying sensible defaults to series volume and query scope
  

Usage is metered on both hourly and daily intervals, making it possible to observe trends and detect gradual changes over time. Enforcement events are logged directly into your telemetry stream, allowing you to query, visualize, and alert on them just like any other operational signal.

For label enforcement specifically, Coralogix now applies a deterministic checksum to oversized label values. This preserves uniqueness and traceability without silently truncating or discarding data, ensuring consistency while remaining within platform constraints.

What happens when a limit is reached

Reaching a fair usage limit does not result in a hard stop.

Instead, the platform provides clear and immediate feedback, including:

• Logs describing which limit was reached and which data was affected
  
• Diagnostic context to help identify the underlying cause
  
• Guidance on next steps, such as data optimization, query refinement, or a usage review
  

This approach ensures that teams remain in control. There is no ambiguity about why behavior changed, and no need to infer the root cause after the fact. When limits need to be adjusted to reflect legitimate usage patterns, account teams are available to help evaluate and apply changes.

Monitoring and alerting on usage behavior

Usage visibility is treated as a first-class observability signal.

Limit enforcement events are emitted to the labs.limitViolations dataset and can be queried using standard Coralogix tooling. Teams can build dashboards, create alerts, or correlate limit activity with other signals across logs, metrics, and traces.

In addition, metrics such as unique_series_daily make it easy to track cardinality growth over time and take proactive action before thresholds are reached. This enables smooth, predictable scaling as telemetry volumes increase.

A measured alternative to hard quotas

Traditional quota systems tend to be abrupt. They block usage without explanation and provide little insight into how limits were reached.

Fair usage limits take a different approach. They are designed to expose system behavior, not obscure it. Teams can see which patterns are contributing to enforcement, understand why those patterns matter, and decide how to respond.

The result is a system that supports growth with guardrails, rather than walls.

Designed to support reliable scaling

Fair usage limits exist to make Coralogix more predictable, transparent, and resilient at scale.

By combining explicit thresholds with actionable feedback and flexible responses, the platform enables teams to operate with confidence as their observability footprint grows.

You can review current usage and applicable limits under Settings > Metric Data > Fair Usage Limits, where limits are grouped by ingestion, metrics, and query behavior. Diagnostic logging can be enabled selectively, depending on the level of visibility required.

For questions or adjustments based on real-world usage, your Coralogix team is available through in-app chat or your account contact.

Scaling observability works best when the system is clear about how it behaves. Fair usage limits are one step toward that clarity.