Tootfinder

Opt-in global Mastodon full text search. Join the index!

@hynek@mastodon.social
2026-03-25 12:13:22

speaking of Sentry: y'all might want to add a "werkzeug!=3.1.7" pin if you run Werkzeug (maybe via Flask) in a cluster with weird hostnames:
github.com/pallets/werkzeug/is

3.1.7 version introduces a BC break in host validation · Issue #3141 · pallets/werkzeug
A feature to validate host http header no matter what introduces backwards compatibility break. The bug presents itself by regarding the host invalid (false-positive) A host like 069ef83c8f81497b8e...

@detondev@social.linux.pizza
2026-04-25 17:58:58

My grandad has a ton of old papers from his field lying around and they're so beautiful

Schematic Of Bioluminescence Sensors On The JSL Submersible DIVING COMPARTMENT 2-CHANNEL SCALAR (PMT CTS/60 sec) 1024 MULTI-CHANNEL ANALYZER (PMT CTS/10ms) ( 1024 MULTI-CHANNEL ANALYZER PMT CTS/10ms) OTRONA MINI-COMPUTER + DUAL DISK DRIVES (DATA STORAGE + CONTROL) JOHNSON SEA LINK SUBMERSIBLE PMT ELECTRONICS PUMPED BIOLUMINESCENCE DETECTOR SEAWATER INTAKE AERYLIC WINDOW SPIE Vol 489 Ocean Optics cs VII (1984)/ 87 ELECTRONICS OPEN BIOLUMINESCENCE DETECTOR PMT DIVING SPHERE BENTH…
SIDE VIEW OF THE NOSC LABORATORY PLANKTON TEST CHAMBER |- 2 7.8 in. PORT CAP PMT HOUSING 452 1900 DIAPHRAGM QUARTZ -1 in.- 3 3/4 in. 8575 PMT COPEPOD FILTER PAPER WATER-LINE LUCITE BASE 8575 PMT BASE CHAMBER VACUUM LINE CHAMBER STAND TO VACUUM FLASK Figure 5
Nitrate Ammonium Light Phytoplankton + Cyanobacteria Chlorophyll RNA DNA Oxygen Bacteria Zooplankton Feces Phaeopigments Sedimenting Matter Loss Light Small Egested Particles Phaeopigments Photo-oxidized Product Nitrate Figure 1. Trophic relationships of the planktonic community of the upper ocean. and ammonium are included because they are considered to be limiting to growth rates and compartment sizes.
Figure 2: Laminar drop splash showing dyed tip drops and vortex rings formed from an earlier splash. Impact velocity 2 m s-1. Figure 3: Bubble and down jet formed from a 3 mm drop at terminal velocity. Original drop dye can be seen beneath the crater. This forms material for the jet tip should the bubble break.
@jamesthebard@social.linux.pizza
2026-03-07 18:11:46

The rewrite of the Sisyphus server continues, I've got all of the GET endpoints and about 60% of all the endpoints finished overall. Some initial testing on the `/workers` and `/queue` endpoints show some worrying results from the old version and some great results from the new version. The `/workers` endpoint started throwing errors at about 800 RPS on the Python/Flask version (the old one), the rewrite is showing around 4800 RPS (Golang/Gin). The `/queue` endpoint doesn't show a…

The graph of the `/workers` endpoint. The left-hand graph is the Python/Flask version (old) and the right-hand graph is Golang/Gin (new)
The graph of the `/queue` endpoint. The left-hand graph is the Python/Flask version (old) and the right-hand graph is Golang/Gin (new)
@mot@chaos.social
2026-04-03 20:54:45

Sind User Sessions unter Python Web Frameworks wie Flask und FastAPI wirklich so beschissen, oder bilde ich mir das nur ein? Entweder es werden Session-Daten Client-seitig (signiert) in Cookies bzw. JWT gespeichert, oder die Server-seitigen Implementierungen brauchen aufgeblähte Datenbanken :(

@jamesthebard@social.linux.pizza
2026-03-08 05:14:59

So, decided to get them rookie `/queue` numbers up so did a quick bit of caching. For a 12-job queue, I went from 240 RPS (Python/Flask) to 680 (Golang/Gin) then to 3400 RPS with some proper caching. The `/workers` endpoint would benefit as well, but I think its performant enough though I may end up adding a bit of caching later.
#golang

The old endpoint (left) versus the new endpoint (right). What's more, the rewritten endpoint still had headroom resource-wise on the VM whereas the old one had maxed out the VM's CPU.