​
Configure Debugging Datasets

Relvy structures its debugging around the concept of datasets. A dataset defines a focused scope of data — such as logs, events, or traces — that Relvy queries from your connected observability backends (like Datadog).

By configuring datasets, you’re telling Relvy:

  • What type of data to query (logs, spans, events)
  • Which subset of that data is relevant (using queries)
  • Which provider and environment the data belongs to

In this step, you’ll configure datasets across three observability signals: Logs, Events, and Traces.

​
Configure Logs Dataset

Logs are a critical part of debugging — they provide granular, timestamped context about what’s happening in your services. To configure a logs dataset:

  1. Expand the section titled “What logs do you investigate?”

  2. Select Data Source, Datadog in this example.

  3. Fill in the following fields:

    • Dataset Name: Choose a name that describes the dataset’s purpose (e.g., application_logs, security_logs).
    • Logs Query: Provide a query that filters relevant logs from your observability backend. This query should isolate logs typically used during investigations (e.g., environment-specific logs or service-level logs).

  4. Click Save to finalize the configuration.

Once saved, the dataset will appear in your list of configured datasets. You can create multiple logs datasets to match different investigation patterns - for example:

  • One dataset for frontend application logs
  • Another for backend service logs
  • A dedicated dataset for security-related logs
  • Environment-specific datasets (prod vs staging)

This flexibility allows you to organize and scope your log data in ways that match your team’s debugging workflows.

​
Configure Events Dataset

Events provide higher-level signals such as deployments, alerts, infrastructure changes, or custom business events. These are useful for correlation and root-cause analysis.

To set up events datasets:

  1. Expand the section “What events do you investigate?”

  2. Select Data Source, Datadog in this example.

  3. Provide:

    • Dataset Name: Assign a descriptive label for clarity.
    • Event Query: Define a query that isolates the events you’re interested in (e.g., deployment events, Kubernetes events, alert notifications).

  4. Click Save for each dataset added.

​
Configure Traces Dataset

Traces give visibility into distributed transactions across services, making them essential for understanding latency, dependencies, and service behavior.

To configure trace datasets:

  1. Expand the section “What traces do you investigate?”

  2. Select Data Source, Datadog in this example.

  3. Set the following:

    • Dataset Name: Use a clear and descriptive name (e.g., all_traces, checkout_flows).
    • Trace Query: Provide a trace query that returns spans relevant to your investigations. This may focus on certain services, environments, or request types.

  4. Click Save to confirm.

​
Why This Matters

Each dataset you define tailors how Relvy:

  • Collects and filters data from your observability provider
  • Provides context-specific suggestions during incidents

Properly scoping your datasets ensures Relvy surfaces the most relevant signals when it matters most, without overwhelming users with unrelated data.