infra
Star
Fork
Watch

Commits on master
hosts/spindle-hel: run nixery workflows in 4 GiB so they fit the per-workflow memory cap
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake: bump tangled
committed
hosts/spindle-hel: install the nixos-large image and keep other repos' workflows at 2 vCPUs and 4 GiB
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake: bump tangled
committed
hosts/spindle-hel: keep up to 100 GB of ci caches on disk
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake,modules/hyd-node: hold hydrant's port through restarts so clients wait instead of getting refused
committed
flake: bump hydrant-mock
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake: remove `tangled-next-dev` input
committed
tf/modules/environment: let each bobbin instance take 500 requests at once
committed
flake: bump tangled
committed
flake: bump tangled
committed
tf/modules/environment: run svfe in cloud run's gen2 environment so calls to bobbin don't stall under load
committed
tf/modules/lb: refuse chrome user agents without http/2 and throttle commit, history, compare and search pages
committed
tf/envs/prod: give svfe 64 threads for dns lookups
committed
tf/modules/{lb,environment,envs}: let pinned file pages through at 20 a second and branch ones at 30, without the capture groups cloud armor refuses
committed
tf/modules/control: let the prod deployment make cloud armor policies
committed
tf/modules/environment: let svfe scale to 8 instances per region
committed
tf/envs/prod: serve the web app's built assets from google's cdn
committed
tf/modules/{lb,environment,envs}: put cloud armor in front of prod's web app, banning ips that flood it and capping files at a commit
committed
tf/modules/{cloud-run-app,environment}: let each svfe instance serve 200 requests at once with 2 GiB of memory
committed
flake: bump tangled
committed
flake: bump tangled
committed
flake: bump tangled
committed
hyd-node: small tweaks
committed
flake: bump tangled
committed
hyd-node: liar foster reports to hydrant
committed
flake: bump tangled
committed
hosts/spindle-hel: read a repo's manifest off the foster knot when its own knot doesn't have it yet
committed
flake: bump tangled
committed
bump hydrant-mock
committed
flake: all services use tangled master
committed
Revert "tf/modules/lb: limit per-IP in svfe"
committed
hosts/mirror-gcp: to 8
committed
hosts/mirror-gcp: lower malloc arena max to 2
committed
hosts/mirror-gcp: add service memory limits
committed
tf/modules/lb: limit per-IP in svfe
committed
restore uncommitted changes
committed
tf/envs/prod: bump hydrant box size
committed
bump hydrant-mock
committed
hosts/spindle-hel: accept repo declarations from the liar pds for accounts that haven't moved yet
committed
flake: bump tangled-next-dev
committed
hosts/knot1: let anyone create repos on knot1
committed
flake: bump tangled-next-dev
committed
tf: add delegation envs
committed
hosts/spindle-hel: let ptr.pet administer the spindle
committed
tf/envs/prod: create new repos on knot1 again until knot.tangled.network exists
committed
hosts/mirror-gcp,flake,tf/envs: let the mirror count profile commits by asking prod's deliberi for each person's emails
committed
flake: bump tangled-next-dev
committed
flake: bump tangled-next-dev
committed
flake: bump tangled-next-dev
committed
tf/envs/prod: run bobbin and svfe in us-east1 and asia-southeast1 too
committed