Iris

Iris is a lightweight API designed to provide seamless, low latency access to official Brazilian economic indicators, including the Selic interest rate and IPCA inflation index.

Local Setup

Use the following make commands:

=============================
Available commands:
=============================
down                      Stop containers
load                      Run load test
up                        Setup containers

Telemetry

Distributed Tracing (Tempo):

• HTTP request/response traces
• Redis command execution spans
• External HTTP client traces (BCB API)
• Background job execution flows

Metrics (Prometheus):

• http_server_request_duration_seconds - Latency histograms
• http_server_request_total - Request counters by status
• process_runtime_dotnet_gc_collections_total - GC performance

Structured Logging (Loki):

• JSON-formatted log output
• Correlation IDs for trace-log linking
• Severity-based filtering

Query Examples

TraceQL - Latency analysis:

{service.name="iris-api"} | duration > 500ms

LogQL - Error correlation:

{service_name="iris-api"} |= "Error" | json | line_format "{{.message}}"

PromQL - Throughput calculation:

sum(rate(http_server_request_duration_seconds_count[5m])) by (http_route)

Contributing

Contributions are welcome! Open issues for bugs, questions, or suggestions. Submit pull requests with new resources or improvements.

Further Results

After fully populating Redis, a load test was executed using k6. The results demonstrate consistent low response times and stable throughput under proposed concurrency.