How maxymfarenyk/tracing-tax works
tracing-tax is a research stand for measuring how trace-sampling strategies affect a checkout microservice system. Locust drives requests through the gateway; orders coordinates inventory reservations and simulated payments, while services export traces through the collector to Jaeger. Scenario configuration selects head, tail, or adaptive sampling. Prometheus and Grafana expose resource and collector metrics. The experiment runner gathers load, metrics, and retained traces into CSV; analysis turns runs into summaries and charts. Documented topology and capabilities are shown alongside source-visible service calls; infrastructure wiring is primarily documented in the README.
Main entry points: app.py Gateway · configs OTel Collector · downstream.py HTTP client helper · prometheus.yml Prometheus
Components 23
Checkout services
- Gateway
services/gateway/app.py - Orders
services/orders/app.py - Inventory
services/inventory/app.py - Payments
services/payments/app.py - HTTP client helper
services/common/downstream.py - Fault injection
services/common/faults.py
Tracing pipeline
- Sampling scenarios
scenarios - Adaptive sampler
services/adaptive_sampler/src/adaptive_sampler/__init__.py - OTel Collector
collector/configs - Jaeger
jaeger/config.yaml
Monitoring
- cAdvisor
docker-compose.yml - Prometheus
monitoring/prometheus.yml - Grafana
monitoring/grafana
Experiment workflow
- Experiment runner
scripts/run_experiment.py - Load profile
load/locustfile.py - Trace verification
scripts/verify_trace.py - Monitoring verification
scripts/verify_monitoring.py - Sampling verification
scripts/verify_sampling.py - Run dataset
results/raw/runs.csv
Results analysis
- Results analyzer
analysis/analyze.py - Summaries and charts
results/analysis
Other components
- Locust load
- Redis
Connections 35
- Locust load → Gateway: sends requests
README.md - Load profile → Gateway: targets
README.md - Gateway → HTTP client helper: calls
services/gateway/app.py - HTTP client helper → Orders: HTTP request
services/orders/app.py - Orders → HTTP client helper: calls
services/orders/app.py - HTTP client helper → Inventory: HTTP request
- HTTP client helper → Payments: HTTP request
- Inventory → Redis: increments stock
services/inventory/app.py - Gateway → Fault injection: injects faults
services/gateway/app.py - Orders → Fault injection: injects faults
services/orders/app.py - Inventory → Fault injection: injects faults
services/inventory/app.py - Payments → Fault injection: injects faults
services/payments/app.py - Sampling scenarios → OTel Collector: selects configuration
README.md - Sampling scenarios → Adaptive sampler: enables strategy
README.md - Gateway → OTel Collector: exports traces
README.md - Orders → OTel Collector: exports traces
README.md - Inventory → OTel Collector: exports traces
README.md - Payments → OTel Collector: exports traces
README.md - OTel Collector → Jaeger: exports traces
README.md - cAdvisor → Prometheus: provides metrics
README.md - OTel Collector → Prometheus: exposes metrics
README.md - Prometheus → Grafana: supplies dashboards
README.md - Experiment runner → Load profile: runs load
scripts/run_experiment.py - Experiment runner → Prometheus: queries metrics
scripts/run_experiment.py - Experiment runner → Jaeger: queries traces
scripts/run_experiment.py - Experiment runner → Run dataset: writes rows
README.md - Results analyzer → Run dataset: reads runs
analysis/analyze.py - Results analyzer → Summaries and charts: writes reports
analysis/analyze.py - Trace verification → Gateway: sends request
scripts/verify_trace.py - Trace verification → Jaeger: queries trace
scripts/verify_trace.py - Monitoring verification → Gateway: sends requests
scripts/verify_monitoring.py - Monitoring verification → Prometheus: checks metrics
scripts/verify_monitoring.py - Sampling verification → Gateway: sends requests
scripts/verify_sampling.py - Sampling verification → Monitoring verification: uses query helpers
scripts/verify_sampling.py - Sampling verification → Sampling scenarios: reads settings
scripts/verify_sampling.py