Comper ships.
We man the cluster.

Comper continuously analyses millions of lines of code to map risk, ownership and software evolution, processing customer repositories around the clock. That makes platform reliability business-critical. We designed and operate the observability platform behind it, so Comper's engineering team stays focused on product delivery.

Client
AI-powered code intelligence platform
Initial engagement
Production Observability Stack
Ongoing service
Observability Care
Environment
Multi-cluster Kubernetes

From manual checks to real-time alerting

BeforeIssues went undetected until they reached users.
NowIssues are flagged before they reach users.
No production telemetry: no metrics, logs or traces
Metrics, logs, traces and profiles from every cluster in one place
No alerting on infrastructure incidents
Real-time, symptom-based alerting routed to the right owner
Container restarts went unnoticed
Latent memory issues surfaced and resolved
No clear ownership of alerting or capacity planning
Monthly reviews of alert quality, retention and cost

From implementation to managed operations

  1. 01

    Implementation

    A production-grade observability platform on Kubernetes. Grafana Alloy provides OpenTelemetry-based collection from every cluster into centralised storage: Prometheus and Thanos for metrics, Loki for logs, Tempo for traces and Pyroscope for continuous profiling, unified in Grafana.

  2. 02

    Validation

    Within weeks, the new visibility surfaced latent memory-pressure issues in a number of services. Previously invisible, they could now be diagnosed and resolved before they affected customers.

  3. 03

    Operations

    We now operate as Comper's SRE function under a monthly retainer, covering alert tuning, platform lifecycle management and monthly reviews of retention and cost.

01Sources

02Collector

03Storage

04Visualisation

Kubernetes clustersRemote ingestion from multiple clusters
Grafana AlloySingle OpenTelemetry agent, full telemetry from every node
MetricsPrometheus + Thanos
LogsLoki
TracesTempo
ProfilesPyroscope
GrafanaA single view for incident response and analysis
Long-term retention in S3 via Thanos, with cost-effective downsampling policies

Designed for

High availability

A resilient telemetry pipeline, built highly available from day one.

Team ownership

Owned by Comper's engineers, operated with us as their SRE function.

Growth

New services and clusters plug in without a rebuild.

Measured impact

Measurement in progress

Mean time to detect (MTTD)

Before: user-reported

Measurement in progress

Alert volume per week

Measurement in progress

Actionable alert rate

Measurement in progress

Latent memory issues resolved

Measurement in progress

Monthly storage cost

Measurement in progress

Incidents detected before customer impact

Figures are sourced from Comper's Grafana environment and published with their approval.

BC helped us out quickly by setting up observability in our Kubernetes clusters. The setup is highly available and Victor was very responsive during the whole process, and also available for quick finetuning and setting up extra alerts after the installation. Highly recommended.
Jouke

SRE capability, without the headcount.

Observability Care starts at €750 per month. Begin with a complimentary health check of your current monitoring setup.

View pricing