Travis Reyburn

Travis Reyburn

@treyburn.dev

Software - Maps - Bikes - Hikes Roll pasta, fix bikes. 🌎 Boulder, CO

321 followers1.5K following188 posts

Travis Reyburn
@treyburn.dev

I think the thing I’m trying to be better about is just being fine with throwing a bunch of unworkable Claude slopcode away and trying again. Turning those learnings into a better initial plan for better results. Just etch-a-sketch it.

Travis Reyburn
@treyburn.dev

I’ve been dabbling into simd with Go and this article is an absolute treasure. What a wealth of information to share!

Michael Stapelberg 🐧🐹😺@zekjur.bsky.social

New blog post: Debian Code Search: Fast TurboPFor with Go SIMD 🥳 This August, I deleted the last cgo dependency in DCS, thanks to Go’s recently introduced SIMD support: michael.stapelberg.ch/posts/2026-0... #golang #debian #simd

michael.stapelberg.chDebian Code Search: Fast TurboPFor with Go SIMDThis August, I accomplished what I wanted for many years: I deleted the last cgo dependency in Debian Code Search! This was made possible by Go’s recently introduced SIMD support, because now we can i...
Travis Reyburn
@treyburn.dev

GCP having a pretty bad outage today + all the ongoing issues we’ve had with them recently just really makes me wanna go on-prem again

Travis Reyburn reposted

Boulder Reporting Lab
@boulderrl.bsky.social

#Boulder’s new comprehensive plan could create capacity for up to 40% more housing, largely through “gentle infill” like duplexes, townhomes and ADUs. What gets built depends on zoning and the market. boulderreportinglab.org/2026/08/27/b...

boulderreportinglab.orgBoulder approves land-use plan that could make room for up to 40% more housingThe update to the Boulder Valley’s land-use map follows two years of public engagement and debate and establishes a new long-term development framework for the region.
Travis Reyburn
@treyburn.dev

Woke up to a ruckus going on outside - thought someone was trying to break into our garage - but nope just a black bear mama and her cubs just annihilating our compost bin

Travis Reyburn reposted

tbabb
@tbabb.bsky.social

my car is throwing an engine trouble code, and I can just directly plug an intelligence into it and it will talk to the car computer and just figure it out the present is unbelievably cyberpunk

Travis Reyburn
@treyburn.dev

Finally saw the Odyssey (in glorious 70mm). I have some thoughts. Namely - why did nobody move the dog off the dung heap??? He’s an old dog - he’s going to catch sepsis!

Travis Reyburn
@treyburn.dev

Man - AI has completely eaten up the clouds capacity. We straight up can’t do a Postgres major version upgrade because GCP doesn’t have the capacity available for a measly 64cpu/256gb mem in their “Enterprise Plus” node pool. Scheduled 2 maintenance windows and failed twice now. Ridiculous!

View full thread
Travis Reyburn
@treyburn.dev

Honestly shook hearing them say this. Does this mean we can’t allocate additional disk capacity in our DBs until they resolve this?! Holy moly this feels like a ticking time bomb!

Travis Reyburn
@treyburn.dev

Man - AI has completely eaten up the clouds capacity. We straight up can’t do a Postgres major version upgrade because GCP doesn’t have the capacity available for a measly 64cpu/256gb mem in their “Enterprise Plus” node pool. Scheduled 2 maintenance windows and failed twice now. Ridiculous!

View full thread
Travis Reyburn
@treyburn.dev

Ok now THIS is crazy. I’m hearing from our GCP rep that the reason we can’t provision any new database instances is because gcp-west1 has completely exhausted its SSD capacity

Travis Reyburn
@treyburn.dev

Man - AI has completely eaten up the clouds capacity. We straight up can’t do a Postgres major version upgrade because GCP doesn’t have the capacity available for a measly 64cpu/256gb mem in their “Enterprise Plus” node pool. Scheduled 2 maintenance windows and failed twice now. Ridiculous!

Travis Reyburn
@treyburn.dev

Wow - apparently Google left us in a broken state where our read replicas are no longer able to replicate. Apparently the only solution there is to blow away our read replicas and re-create them. Except now that we’ve done that - GCP still doesn’t have any resources.

Travis Reyburn
@treyburn.dev

So now we can’t recreate our read replicas. Which don’t even need much resources in the first place! 32cpu and 128mem in 2 regions is all we’re asking for. We even have those resources back! But now they’ve already been allocated because everything is so oversubscribed.

View full thread
View full thread
View full thread
Travis Reyburn
@treyburn.dev

Had a fun time and learned a lot on this project! - Nushell was brand new and is gonna be a staple in my toolbox. - Mise! Idk how I’ve gone this long without using this. - Bubbletea continues to be a joy to work with. - GitHub immutable releases continue to trip me up ever new project 🫠

Travis Reyburn
@treyburn.dev

First week of school is over - but that sudden back-to-school micro-biome shift has already gotten my partner sick. T-minus 2 days for me! 😑

Travis Reyburn
@treyburn.dev

Oh woah - Unraid is rebasing onto uCore!? Well I guess I can shelve my bespoke homelab distro project now - and re-focus on my bespoke dev desktop bluefin+dms distro project instead 😂

Travis Reyburn
@treyburn.dev

Incidents are no fun - but I think I’ve learned more about Postgres and query optimizations in the past 3 weeks than the rest of my career combined.

View full thread
Travis Reyburn
@treyburn.dev

Pretty stoked out outcomes from this (and continuing efforts!). CPU on our read replica down from an average of 50% to sub 10%. P95 (read only) query times down from ~150ms to 0.9ms (!!!). Query timeouts clamped down from 10s to 2s.

Travis Reyburn
@treyburn.dev

Incidents are no fun - but I think I’ve learned more about Postgres and query optimizations in the past 3 weeks than the rest of my career combined.

View full thread
Travis Reyburn
@treyburn.dev

If you lateral join on a subquery depending on a WHERE filter - then you better hoist that subquery out or Postgres will just read out every possible value

Travis Reyburn
@treyburn.dev

Beware analyze! If your data is non-homogeneous then you’re doing a coin flip on how good your pg_stats will be. ALTER TABLE foo ALTER COLUMN field SET STATISTICS {between 100 (default) and 10k}. You can trade analyze runtime (defaults to fast) for substantially larger sample sizes for your stats

Travis Reyburn
@treyburn.dev

Incidents are no fun - but I think I’ve learned more about Postgres and query optimizations in the past 3 weeks than the rest of my career combined.

View full thread
Travis Reyburn
@treyburn.dev

ST_Intersects(geometry, geometry) will gladly give you a distance of 0 if you ask for it … after checking every possible point combination?!

Travis Reyburn
@treyburn.dev

Incidents are no fun - but I think I’ve learned more about Postgres and query optimizations in the past 3 weeks than the rest of my career combined.

Travis Reyburn
@treyburn.dev

Forgejo needs a mascot or some kind of cutesy icon. I’m ready to start running an influence campaign on slack to get my company off GitHub 🫠

Travis Reyburn
@treyburn.dev

I think GitHub’s refusal to acknowledge non-catastrophic outages and issues on their status page is really the worst part of their unreliability. I’m sympathetic to the strain their system with this scale. But you gotta be honest and acknowledge it publicly to your users.

Travis Reyburn
@treyburn.dev

Tailscale rules. These guys paid the expert maintainers to investigate and fix an old issue, then funded an open source project to make that kind of debugging easier for the community. Refreshingly level headed leadership. tailscale.com/blog/sqlite-...

tailscale.comHow Tailscale helped find the SQLite WAL-Reset bugTailscale and SQLite developers traced maddening corruption incidents to find the WAL-Reset data race, then uncovered a second stale expression index bug.