Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Polaris super consumer #450

Merged
merged 34 commits into from
Apr 21, 2022
Merged

Polaris super consumer #450

merged 34 commits into from
Apr 21, 2022

Conversation

elaubsch
Copy link
Member

@elaubsch elaubsch commented Mar 30, 2022

This PR creates the Polaris super consumer. The Polaris queue and consumer previously handled spot detection tasks but have been refactored to the spots queue and consumer. The changes to maximum batch size are an attempt to address #449, but the issue is persisting. A temporary solution was implemented in this PR.

willgraf and others added 30 commits April 6, 2021 16:42
* Update consumers to 0.12.0

* Update models for tracking consumer and add TRACK_LENGTH.

* Update chart with new version and new env vars.

* Bump version in all helmfiles too

* Update Chart version too.
* Update autoscaler to 0.7.0 to support kubernetes API 18.x

* Update janitor to 0.6.0 to support kubernetes API 18.x

* Update default KUBERNETES_VERSION to 1.19

This is the earliest supported version on GKE.
* Update app version to 0.8.1

* Add default REACT_APP environment variables

* Bump chart version

* Use new chart version in helmfile
* Pin mistune<2 to resolve issue with m2r and mistune.

* Pin docutils to 0.16 to match deepcell environment.
Uses tensorflow-serving 2.5.3.
* Upgrade default KUBERNETES_VERSION to 1.20.

* Update autoscaler and janitor to latest version with kube 20.x support.

* Upgrade bucket monitor to latest version using Python 3.8
* Update default KUBERNETES_VERSION to 1.21.

* Upgrade autoscaler to use the kubernetes 21 client.

* Upgrade janitor to use the kubernetes 21 client.
This has a patch for the caliban refactor.
Copy link
Member

@msschwartz21 msschwartz21 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you add a comment to the PR about why TF_MAX_BATCH_SIZE had to be reduced so that we have a record somewhere? Otherwise LGTM!

@msschwartz21 msschwartz21 added the enhancement New feature or request label Apr 21, 2022
@msschwartz21 msschwartz21 merged commit 7fceb7d into stable Apr 21, 2022
@msschwartz21 msschwartz21 deleted the polaris-super-consumer branch April 21, 2022 23:54
msschwartz21 added a commit that referenced this pull request Apr 22, 2022
* Adding Polaris consumer and Polaris zip consumer (#448)

* Polaris super consumer (#450)

* Correct integration-test to use google-github-actions/setup-gcloud@main instead of @master

Co-authored-by: Emily Laubscher <43769564+elaubsch@users.noreply.github.com>
willgraf added a commit that referenced this pull request Jul 1, 2022
* Adding Polaris consumer and Polaris zip consumer (#448)

* Initial commit of Polaris consumer

* Add polaris autoscaling

* Update spot detection model

* Update Docker image tags frontend 0.9.0 and consumer 0.14.0

* Polaris super consumer (#450)

* Initial commit of super consumer

* Add changes from stable

* Initial commit of spot zip consumer

* Revert React setting to stable

* Decrease TF_MAX_BATCH_SIZE

* Pin Jinja version

* Update Docker tags and pull policies

* Correct integration-test to use google-github-actions/setup-gcloud@main instead of @master

* chore: update to Kubernetes 1.22 (#461)

* chore: update default k8s version

* fix: update frontend ingress for v1.

Followed guide:
https://kubernetes.io/docs/reference/using-api/deprecation-guide/#ingress-v122

* chore: update cert-manager to v1.8.2.

* chore: update rbac auth API to v1.

* chore: use apps/v1 instead of extensions/v1beta1.

* chore: update prometheus-adapter to newest v2.x

* chore: bump frontend chart version

* chore: update prometheus-redis-exporter to latest 3.x

* chore: bump frontend default container version

* fix: update service for frontend.

* chore: update ingress to latest version.

* feat: upgrade prometheus-operator to kube-prometheus-stack.

Lots of major version changes, but that is because of the CRDs.
Seemingly no values change is required.

* fix: update prometheus URL.

* chore: update redis exporter to latest version.

* fix: update make cmd for new prometheus name

* chore: bump prometheus-adapter to latest version

* chore: rename prometheus-operator to kube-prometheus-stack

* chore: update redis to latest helm chart

* feat: remove service from redis-consumer chart.

bump chart version and update default app version to 0.15.0

* chore: bump version to 0.16.2 (prod version)

* fix: update display IP command.

* chore: update docs with minor OVPN note.

* fix: spot-consumer custom metric name.

* chore: update autoscaler and janitor to k8s client 22.x

* chore: bump version to 1.8.0

* Merge several updates from dev branch into stable (#463)

* Bump Mesmer model version to 8 and update consumer to 0.16.1 (#455)

Co-authored-by: William Graf <7930703+willgraf@users.noreply.github.com>

Co-authored-by: Emily Laubscher <43769564+elaubsch@users.noreply.github.com>
Co-authored-by: msschwartz21 <msschwartz21@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants