Stephen Wicklund issues

Results 189 issues of


                                            Stephen Wicklund

Improve performance of live-query system further, eg. by avoiding more calls to clone()

Fix that using cranelift + pyroscope together, causes seg-faults during heavy activities, eg. large map import [once complete, probably enable both again]

Fix that the weekly backup process (using pgdump) can result in "dangling" db connections, which stay alive forever (show as "active" in session-manager in dbeaver), and which prevent schema changes and such until terminated

I do not currently know what causes the pgdump connections to become "dangling" like this... perhaps it's due to a crash of the app-server while the SQL query in the...

Fix whatever is causing the pg_wal folder to grow so large in prod cluster!

For now, when wal folder grows too large, use following fix route: * 1\) Increase size of the out-of-space PVC(s), using Lens. (instead of default `10000Mi`, set to `10100Mi` or...

Extract link-preserver system from app-server into rust-shared package, then implement it for monitor-backend as well (and use to fix monitor-frontend's issue of websocket disconnecting on homepage)

Set up GitHub Action that will do a fresh compile of the rust "app-server" crate on each commit, and record the "--timings" html output into "debate-map/compile-timings" repo

How to have Rust record compile-timings info: https://doc.rust-lang.org/cargo/reference/timings.html GitHub actions docs (entry point): https://docs.github.com/en/actions

Fix that pgdump backup route is broken atm, due to nginx timing out requests before they are established/solidified

Two routes to solve this: 1) Increase nginx timeout somehow; equal to issue #320. (this is not trivial, as this is a lower level config option that is not exposed...

Expand web-vcore's SyncDepsToOuterProject.js script to also handle ".yalc" folder local-overrides (as much as possible)

MS if app-server crashes due to OOM (or whatever caused issue #335), some indication of crash type is at least logged (if possible)

Basic steps: * Create a test-case somewhere that causes app-server's memory usage to exceed the memory limit (I guess, just intentionally create a large memory leak that keeps growing), and...

Re-add UI testing system. (eg. add cypress package back, figure out how to integrate with CI without sacrificing ui-centric testing, etc.)

Note: Maybe have this testing system just be part of the load-testing system. (or if not, then decide on the boundary between the two)

frontend