Skip to main content
Gitaly provides comprehensive monitoring and observability capabilities to help you track performance, identify bottlenecks, and troubleshoot issues.

​
Grafana Dashboards

Gitaly includes pre-built Grafana dashboards for monitoring performance on GitLab.com and self-managed instances.

​
Overview Dashboard

The main overview dashboard provides high-level metrics across all Gitaly operations:
  • Overall RPC request rates
  • Latency percentiles
  • Error rates
  • Resource utilization
View Gitaly Overview Dashboard

​
Feature Request Rate Overview

For detailed analysis of specific features and RPC methods: View Features Overview Dashboard

​
Editing Dashboards

Use the Grafana web interface to customize dashboards:
Always click the ‘Save’ button at the top of the Grafana screen when making changes.

​
Tiled (Repeated) Dashboards

When editing tiled dashboards:
  • Edit only the first tile (top left)
  • Settings from the first tile automatically apply to all other tiles
  • Changes to any other tile will be lost
Dashboard drop-down menus (for GRPC method names, Prometheus jobs, etc.) use Prometheus queries:
  1. Click the gear icon at the top of the dashboard
  2. Navigate to the ‘Templating’ section
  3. Edit the Prometheus queries as needed
Grafana templates use a combination of PromQL and Grafana-specific modifiers.

​
Prometheus Metrics

Gitaly exposes metrics via Prometheus for monitoring:
  • RPC request rates and latencies
  • Git command execution times
  • Repository access patterns
  • Resource consumption (CPU, memory, disk I/O)
  • Concurrency and rate limiting events

​
Backpressure Metrics

Metrics provide visibility into how concurrency and rate limits are being applied. See the backpressure configuration page for details on limit settings. For comprehensive metric details, refer to the GitLab Documentation.

​
Ad-hoc Latency Analysis with ELK

For more accurate latency percentiles than Prometheus exponential buckets provide:
  1. Go to ELK
  2. Click ‘Visualize’
  3. Search for gitaly rpc latency example
  4. Edit the query as needed for your analysis
Use ELK for detailed percentile analysis when Prometheus bucket sizes don’t provide sufficient granularity.

​
Distributed Tracing

Gitaly supports distributed tracing through LabKit using OpenTracing APIs.

​
Enabling Tracing

By default, no tracing implementation is linked. Enable tracing by compiling with build tags: 
make BUILD_TAGS="tracer_static tracer_static_jaeger"

​
Supported Providers

  • Jaeger - Recommended for most deployments
  • Other OpenTracing-compatible providers via LabKit

​
Configuration

Configure tracing via the GITLAB_TRACING environment variable: 
GITLAB_TRACING=opentracing://jaeger ./gitaly config.toml

​
Continuous Profiling

Gitaly supports continuous profiling through LabKit using Stackdriver Profiler. For setup instructions, see the LabKit monitoring documentation.

​
Performance on GitLab.com

To see real-time performance metrics from GitLab.com production: These dashboards show how Gitaly performs at scale with millions of repositories and users.

Build docs developers (and LLMs) love

Get started for freeTalk to us