Hacker News Reader

Comment by amelius | Hacker News Reader

What kind of "tracing" is this about?

Distributed tracing, a way of tracing requests across services, e.g. Zipkin [0] (which is what I believe Twitter uses), Jaeger [1] or Lightstep [2]

[0] https://zipkin.io/ [1] https://www.jaegertracing.io/docs/1.18/architecture/ [2] https://lightstep.com/distributed-tracing/

Also https://opentracing.io/

Now merged with OpenCensus to form https://opentelemetry.io

Think of it like a stack trace, except across processes.

You have to instrument your code and libraries to pass along a context that gets picked up by each piece along the call tree, and then each piece gets reported to a collector.

It's useful for microservices, and even for non-distributed web services if you have heavyweight libraries that it must interact with (databases, caches, queues, etc).

It only really works if all of the pieces understand each other, thus the OpenTelemetry [1] initiative.

[1] https://opentelemetry.io/

I think... the kind where you assemble a beginning to end view of a request and all the micro services it hits by crawling through log files.

first web world problems

I interpret this as meaning that this is only a challenge for the kind of over-scale systems that only a handful of companies have.

You could be a pretty significant online retailer (Asos, say), run everything on a few instances of a monolith, and have all your information in easily interpretable log files, without needing to reach for distributed tracing tools.

Yeah, this tends to be more of an issue for consumer-grade internet services (where the majority of users are not customers).

Again, these tools are incredibly useful when you have a distributed system, and kinda pointless otherwise.

Not really log files as I'm familiar with them. More structured. At least in the implementation I'm familiar with.

This seems really esoteric to distributed systems...

Yeah i guess. I think it's that it's more useful in distributed systems, and most people don't need it.

But when you do need it, you really need it.

Might be best to dive into the literature starting here:

Magpie: Online Modelling and Performance-aware Systems

"Understanding the performance of distributed systems requires correlation of thousands of interactions between numerous components ..."

https://pdfs.semanticscholar.org/cb76/9b0c983bce6a73cafd1f8a...

Reply on news.ycombinator.com