hussein-awala commented on code in PR #59: URL: https://github.com/apache/airflow-ci-infra/pull/59#discussion_r1741277701
########## helm/values/gha-runner-scale-sets/runners.yaml: ########## @@ -0,0 +1,49 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +--- +image: ghcr.io/apache/airflow-ci-infra/actions-runner:20240830-rc1 +runnerScaleSets: + arc-small-amd: + minRunners: 0 + maxRunners: 30 + size: small + arch: x64 + arc-medium-amd: + minRunners: 0 + maxRunners: 30 + size: medium + arch: x64 + arc-large-amd: + minRunners: 0 + maxRunners: 30 + size: large + arch: x64 + arc-small-arm: + minRunners: 0 + maxRunners: 30 + size: small + arch: arm64 + arc-medium-arm: + minRunners: 0 + maxRunners: 30 + size: medium + arch: arm64 + arc-large-arm: + minRunners: 0 + maxRunners: 30 + size: large + arch: arm64 Review Comment: > This totals up to possible a large number of runners -- I worry about the possible cost given we don't have infinite credits. Running 30 nodes for 3h costs the same as running 90 nodes for 1h. I don't say that we have to scale to 1k, but we shouldn't worry about the max number of nodes/pods, we just need to correctly configure the scaling of ARC and nodes. > And we will need far less ARM runners - they are currently disabled, but unless we want to enable tests for ARM (whcih we could now) we need them only for image building. In the first phase, we will only use AMD nodes/runners, but I plan to switch all the CI jobs to ARM runners (except the AMD docker image build), where ARM nodes are faster and cheaper. For example, for a node with 32Gb of RAM and 8 VCPUs: - m7a.2xlarge --> 0.1698$/h - m7g.2xlarge --> 0.1202$/h (30% cheaper) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
