zzstoatzz.io
logfire client for zig
use BatchSpanProcessor for async export
BasicSpanProcessor exports synchronously on every span.end(), causing
major latency overhead (~100-200ms per span for OTLP HTTP requests).
BatchSpanProcessor queues spans and exports them in batches every 1s
via a background thread, eliminating export latency from the request path.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
+6
-3
src/otel_wrapper.zig
+6
-3
src/otel_wrapper.zig
···
119
119
otel_api.trace.Sampler{ .keep = {} },
120
120
);
121
121
122
-
// add OTLP exporter pipeline using pipelineBuilder pattern
122
+
// add OTLP exporter pipeline using BatchSpanProcessor for async export
123
+
// (BasicSpanProcessor exports synchronously on every span.end(), causing latency)
123
124
var builder = provider_ptr.pipelineBuilder();
124
125
builder = builder.with(
125
-
otel_sdk.trace.BasicSpanProcessor.PipelineStep.init({})
126
-
.flowTo(otel_exporters.otlp.OtlpTraceExporter.PipelineStep.init(otlp_config)),
126
+
otel_sdk.trace.BatchSpanProcessor.PipelineStep.init(.{
127
+
.export_interval_ms = 1000, // export every 1s
128
+
.max_queue_size = 2048,
129
+
}).flowTo(otel_exporters.otlp.OtlpTraceExporter.PipelineStep.init(otlp_config)),
127
130
);
128
131
try builder.done();
129
132