(this is one of the bigger opportunities for node competitors who will remain unnamed)https://twitter.com/mipsytipsy/status/1222739530916102144 …
-
Show this thread
-
to expand: this sort of observability is implemented on top of async_hooks in node, which was a gargantuan effort, but ultimately came too late for that platform — the ecosystem was built around different assumptions.
1 reply 0 retweets 4 likesShow this thread -
Replying to @isntitvacant
Yes please! Did some research about this in June for uhh, another language. Part of what came out of it was async-log that adds event loop context to log messages: https://docs.rs/async-log/2.0.0/async_log/ … The other thing missing is probe support; structured hooks for dtrace/bpftrace would be
.1 reply 0 retweets 1 like
What I'd love to see for Rust specifically is to: 1. integrate the log crate into the stdlib 2. add log support throughout the API. Since the log crate is implemented through visitors, this would allow adding any form of {probes,breakpoints,delays} to anywhere in stdlib.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
basically this.
Your process-level debugger no longer works now that functions hop the network. Observability solves this by packing context up with the request as it hops from process to process, forcing an fsync after every successful hop.