Monitoring Go Applications with OpenTelemetry

OpenTelemetry is a CNCF sandbox project for standardizing application tracing and monitoring across multiple programming languages, platforms and monitoring vendors. This talk provides a brief introduction to OpenTelemetry, explores the OpenTelemetry Go library and demonstrates how it can be used to make Go applications observable.

Transcript

1. GoDays Berlin | 22.01.20
2. Distributed tracing is great! (...then why don’t we see it everywhere? )
3. https://xkcd.com/927/
4. Johannes Liebermann Software Developer, Kinvolk Github: johananl Twitter: @j_lieb Email: johannes@kinvolk.io
5. The Kubernetes Linux Experts Engineering services and products for Kubernetes, containers, process management and Linux user-space + kernel Blog: kinvolk.io/blog Github: kinvolk Twitter: kinvolkio Email: hello@kinvolk.io
6. • Distributed tracing in 30 seconds • Main challenges with distributed tracing • Introduction to OpenTelemetry • Demo — instrumenting a distributed Go application • A peek into the OpenTelemetry Go library
7. This talk assumes familiarity with distributed tracing.
8. helloHandler := func(w http.ResponseWriter, req *http.Request) { ... ctx, span := tr.Start( req.Context(), "handle-hello-request", ... ) defer span.End() _, _ = io.WriteString(w, "Hello, world!\n") }
10. • A span measures a unit of work in a service • A trace combines multiple spans together handle-http-request query-database render-response 5ms 3ms 2ms trace span
11. Cool!
12. So why doesn’t everybody use it?
13. • Instrumentation == code changes • Hard to justify reducing team velocity for tracing • You can’t have “instrumentation holes” ◦ At the very least you must propagate context
14. • Vendor-locking for tracing is especially problematic • Importing a vendor-specific library is scary ◦ What if my monitoring vendor raises prices? • Open-source libraries must remain neutral ◦ You can’t require users to use a specific vendor ◦ Maintaining support for multiple vendors is a lot of work
15. • Does distributed tracing conflict with microservices?
16. • Multiple microservices • Multiple programming languages and frameworks • Multiple protocols (HTTP, gRPC, messaging, ...) • Multiple tracing backends (Jaeger, Zipkin, Datadog, LightStep, NewRelic, Dynatrace, …)
17. Is there a solution?
18. https://xkcd.com/927/
19. Standards!
20. Lack of standards is especially costly for distributed tracing.
22. OpenCensus
23. • A CNCF project which standardizes tracing APIs • API only — no implementation! • Released in December 2016 • Notable contributors: ◦ LightStep, Uber, Instana, SolarWinds, NewRelic, Datadog, Red Hat, ... • Supported by: ◦ LightStep, Datadog, Instana, Dynatrace, Jaeger, ...
24. • A Google project which has been open-sourced • API and implementation • Released in January 2018 • Notable contributors: ◦ Google, Microsoft, Splunk, Honeycomb, SignalFx, ... • Supported by: ◦ Stackdriver, Datadog, Honeycomb, AWS X-Ray, Zipkin, ... OpenCensus
25. OpenCensus ⊻
26. + = May 2019
27. • Announced May 2019 • The next major version of both OpenTracing and OpenCensus • A real community effort • A spec and a set of libraries • API and implementation • Tracing and metrics
28. • API ◦ Follows the OpenTelemetry specification ◦ Can be used without an implementation • SDK ◦ A ready-to-use implementation ◦ Alternative implementations are supported • Exporters • Bridges https://github.com/open-telemetry/opentelemetry-specification/blob/master/specification/library-guidelines.md
29. • Library developers depend only on the API • Application developers depend on the API and on an implementation • Monitoring vendors maintain their own exporters
30. • I may want to use an instrumented 3rd-party library without using OpenTelemetry ◦ If no implementation is plugged in, telemetry data is not produced • My code should not be broken by instrumentation ◦ The API package is self-sufficient thanks to a built-in noop implementation • Performance impact should be minimal ◦ No blocking of end-user application by default ◦ Noop implementation produces negligible overhead ◦ Asynchronous exporting of telemetry data
31. • Current status: pre-release • Production readiness: 2nd half of 2020 • Libraries for: Go, Python, Java, C++, Rust, PHP, Ruby, .NET, JavaScript, Erlang
32. Latest release: v0.2.1 (alpha) ✓ API (tracing + metrics) ✓ SDK (tracing + metrics) ✓ Context propagation ✓ Exporters: Jaeger, Stackdriver, Prometheus (metrics) ✓ OpenTracing bridge
33. How do I instrument my Go services?
34. // Explicit span creation. handler := func(w http.ResponseWriter, r *http.Request) { ctx, span := tr.Start(r.Context(), "handle-request") defer span.End() // Handle HTTP request. } // Implicit span creation. err := tr.WithSpan(ctx, "do-stuff", func(context.Context) error { return do_stuff() } )
35. // Log an event on the span. span.AddEvent(ctx, "Generating response", key.New("response").String("stuff") ) // Set key-value pairs on the span. span.SetAttributes( key.New("cool").Bool(true), key.New("usefulness").String("very"), )
36. • “Context” refers to request-scoped data ◦ Example: request/transaction ID • Context is propagated across a request’s path • Distributed tracing relies on context for span correlation ◦ Trace ID and span ID must be propagated • Two types of context propagation: in-process and distributed
37. • Distributed context propagation is protocol-dependent! // Inject tracing metadata on outgoing requests. grpctrace.Inject(ctx, &metadata) // Extract tracing metadata on incoming requests. metadata, spanCtx := grpctrace.Extract(ctx, &metadataCopy)
38. Frontend Seniority Field Role HTTP gRPC
39. Demo
40. Let’s peek into the OpenTelemetry Go library.
41. handleIncomingRequest() { span := tr.Start() defer span.End() library1.doSomething() } library1.doSomething() { // Current span? library2.doSomething() } library2.doSomething() { // Current span? } User Application 3rd-party library 3rd-party library
42. • Used among functions or goroutines within a service • Must be thread-safe • Two main approaches: ◦ Implicit — thread-local storage, global variables, … ◦ Explicit — as an argument in function calls
43. • https://golang.org/pkg/context/ • Commonly employed for cascading request cancellation • Can be used for propagating request-scoped data • Thread-safe • You may already be using context anyway
44. // Set current span. func ContextWithSpan(ctx context.Context, span Span) context.Context { return context.WithValue(ctx, currentSpanKey, span) } // Get current span. func SpanFromContext(ctx context.Context) Span { if span, has := ctx.Value(currentSpanKey).(Span); has { return span } return NoopSpan{} } api/trace/current.go
45. Conclusion
46. • It’s a lot of work ◦ Hopefully you’ll never have to re-instrument your entire codebase ◦ Auto-instrumentation is in the works • We can’t vendor-lock ◦ Architecture encourages separation of concerns • Distributed tracing vs. microservices ◦ A good balance between freedom and uniformity
47. Johannes Liebermann Github: johananl Twitter: @j_lieb Email: johannes@kinvolk.io Kinvolk Blog: kinvolk.io/blog Github: kinvolk Twitter: kinvolkio Email: hello@kinvolk.io