[question] What's the overhead of having spy.create() in production?

Trying to decide whether to enable rxjs-spy in production. Ideally we would like to have it there in order to inspect our streams, but it doesn’t say much in the docs about how significant the overhead is.

The only code that would run in production is spy.create() and tag() on streams, but as far as I understand rxjs-spy patches Observable.subscribe, so perhaps this isn’t very desirable.

Any thoughts on this? 😃

We’ve tried exposing spy.create in production, but the issue then is that if we create the spy later, we lose out on tagging a lot of streams that have already been created.

[cartant]

It’s not mentioned because it’s not at all recommended. When it’s enabled/created everything runs through it. Everything. tag operators have essentially no overhead - as they are little more than annotations - but you definitely do not want to create a spy in a production environment.

Anyway, rxjs-spy is going to be deprecated soon and with its replacement - the RxJS DevTools I’m working on - you will be able to do what you want in a production environment with no overhead if the DevTools are not installed.

[cartant]

… I’m working on

… in a private repo.

I’ll be presenting a talk on the DevTools at RxJS Live - in Las Vegas - in September. Expect more information regarding the tools - and a beta release - to be made available around the time of the talk.