Envision

Links

Docs

Explore

Docs

Infrascructure Components

Infrastructure Diagram:

⁠

flask-backend cloud run (repo:flask-backend)

Face-detection cloud run (repo:pre-processor)

⁠

duplicate-images-v2-prod⁠

cloud function attached to buckets events of envison-us-central1-prod bucket

envision_endpoint_a1111_pubsub (id: w7gv1tqn0cgi4h) Runpod Endpoint for inference jobs

⁠

generate-signed-url⁠

cloud function (not in use)

⁠

notify_slack⁠

cloud function for billing account alerts

⁠

jetrr-cdn⁠

CDN for jetrr buckets(not in use)

load balancers for cdn

⁠

jetrr-cdn-load-balancer⁠

⁠

jetrr-cdn-load-balancer-production⁠

⁠

custom kubernetes job logger deployment (repo:k8s-event-streamer-service)(not in use)

gke-job-status-update-gpu-cluster-auto log sink for logging (in cloud console for log routers)(not in use)

daemonset for caching training image in cpu nodes

image-prepull-daemonset.yml⁠

(repo: kubernetes-client)(not in use)

kubernetes cluster gpu_cluster_auto:(not in use)

image caching is enabled

basic node pool with 1 cpu node for image caching and log streaming service

default node pool for gpu nodes for running stable diffusion jobs

taints

⁠

nvidia.com/gpu:NO_SCHEDULE⁠

⁠

temporary: NO_SCHEDULE

cloud schedulers

⁠

envision-failure-recovery-system⁠

⁠

pub-sub-topics and subscriptions (description in terraform and kubernetes-client repo)

⁠

fcm_notifications_topic⁠

⁠

fcm_notifications_topic-sub⁠

⁠

gke-job-status-update⁠

⁠

gke-job-dev-backend⁠

⁠

images-pre-processed⁠

⁠

post-face-detection-handling-subscription⁠

⁠

uploaded-images⁠

⁠

image-cropper-subscription⁠

⁠

notify-slack⁠

#This topic is not used by envision backend, it’s used for jetrr-cloud project billing alerts

⁠

gcf-notify_slack-us-central1-notify-slack⁠

⁠

dead-letter-topic (to be created)

dead-letter-sub (to be created)

Slack channels

ios-crashes-dev

api timeout notification channel dev

failure revovery job summary

admin panel

dev

prod

Deployment strategy for Production

update configs-prod according to configs dev i.e. update configs collection in firestore

update pre-processor

update backend

check fcm, slack and other webhooks

update runpod prod endpoint

update and test cloud scheduler for Failure Recovery Cron Job

update and Test

duplicate-images-v3-prod⁠

cloud function for replication of data (especially when updating infra with terraform because terraform resets some permissions which causes the function to stop working and you have to manually give the cloud function that permission back)

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.