Benjamin Coe

I've spent *maybe too much* time thinking about logging.

At npm, logging was my main tool for debugging the registry. At Google, I learned to log proactively so debugging rarely required a deployment.

Now, as a PM at Sentry, I get to learn how developers across many companies think about logging.

I've collected my favourite lessons:

blog.sentry.io/logging-best-p…

argument: dashboards are still applicable in a world where analysis of telemetry is being done with agents. they're just more ephemeral -- perhaps conveying info from a single investigation -- and *way* easier to build.

blog.sentry.io/dashboard-upda…

Good sign: we immediately relied on Application Metrics @sentry while building them.

They're launching today 👏

Why care?
• dead simple to instrument
• connected to spans, errors, logs → actually see why something’s off the rails
• generous free tier — great for hobby projects

blog.sentry.io/introducing-ap…

I'm having far too much fun stress testing @sentry's new Application Metrics product.

I can't decide if I like the "scrum_token" or "disruption_id" more.

my team @sentry just released a feature that lets you
create killer dashboards for your app with an AI agent 🤖

it honestly feels a lot like magic -- see my example screenshots.

If you use @sentry , try the feature out! feedback is appreciated 👏

github.com/getsentry/sent…

used @claudeai to help fix cross-platform tests, getting out a fix for a CVE in a transitive dependency.

This would have taken me many hours of work in the past spinning up a VM, etc.

github.com/istanbuljs/nyc…

the future is cool.

made an extremely niche app with @claudeai this weekend for measuring the angle of adjustable climbing walls...

I have to admit, a future where you can whip together something like this *in a morning*, for a very specific use case, is pretty compelling.

working on productionizing your awesome new product? @NikolovLazar and I are putting together a series of posts about using @sentry to identify, and fix, common scaling problems:

blog.sentry.io/paginating-lar…

Not far enough. IMO, you're ngmi if you sleep.

just shipped a feature at @Sentry that notifies you of web vitals performance issues... and, more importantly, it can root cause and fix 'em for you.

as a self proclaimed terrible frontend developer, I find this pretty cool.

blog.sentry.io/meet-web-vital…

I built a retro adventure game for @getsentry hack week.

You should try it out: telemetry-quest.vercel.app

Note, it doesn't work great on mobile yet.

had the realization this weekend that LLMs have gotten good enough for me to rapidly prototype snarky Sierra-style games.

I was thinking of this incident at @npmjs when I helped build Sentry's cache monitoring.

I love checking my dashboard, but it's rad.

Jokes aside, I'm really excited to see my team's work at @getsentry coming together. If you build software, you should check it out.

Just released v18.0.0 of yargs. It's ESM first 🏄 (no longer shipping CJS) and requires Node >= 20.19.0.

github.com/yargs/yargs/bl…

a Pop art version of Finn, post bath, with ChatGPT.

I think I'm starting to understand the appeal of this ML thing.

In the newest versions of Node.js, did you know you can now require ESM modules?

This has motivated me to do my first major overhaul of yargs in years. It will soon be ESM first:

github.com/yargs/yargs/pu…

Idiocracy is my favourite documentary.

imagine you could do anything...

what metrics do you wish you could collect to help understand and improve your production frontend app?

some ideas:

- component-level performance measurements.
- a central collector for the User Timing API.
- something, something, LLMs.

These CAPTCHAs are getting too hard.