this presentation are personal to the author of each respective talk and do not represent the views or opinions of any organization unless explicitly stated. All content provided on this presentation is for informational purposes only.
◦ →e.g.) 2019/8 Service disruption at AWS Tokyo region ◦ Multiple regions, compliance with laws etc. ◦ Want to use Google Cloud Spanner (GCP) ◦ (Full-managed) Cloud Run GA’ed (Jul 2019): https://medium.com/google-cloud-jp/cloud-run-ga-fb31378cd0a1 • →cloud native ◦ →Kubernetes？ ▪ →Kubernetes specialist required, hard Ops • →Extract the common code as framework!!! 2019/8のAWS Tokyo障害: https://aws.amazon.com/jp/message/56489/
production: stage: deploy before_script: - npm config set prefix /usr/local - npm install -g serverless script: - serverless deploy --stage production --verbose environment: production ↑ Add this to your .gitlab-ci.yml to run CI pipeline workﬂow. https://gitlab.com/gitlab-org/serverless/examples/serverless-framework-js/blob/c4bab3616b0ccea96c88d8a28a1ca934ff55e0f8/.gitlab-ci.yml with Serverless Framework
.serverless:build:functions environment: production functions:deploy: extends: .serverless:deploy:functions environment: production ↑ Add this to your .gitlab-ci.yml to run CI pipeline workﬂow. https://gitlab.com/knative-examples/functions/blob/2741e54eb82f882179114590df72cb73074d1c48/.gitlab-ci.yml with Knative (thru gitlabktl) GitLab / OpenFaaS runtimes
.serverless:build:image deploy: extends: .serverless:deploy:image ↑ Add this to your .gitlab-ci.yml to run CI pipeline workﬂow. https://gitlab.com/knative-examples/knative-ruby-app/blob/ece26c9a98eb0c2cafb70c7904c9dbc35f0a0ded/.gitlab-ci.yml with Knative (thru gitlabktl) Your Dockerﬁle required
as important as appsec on non-serverless ◦ 4 of top 6 attacks are web-application securities • Higher responsibility on serverless with shared responsibility model • For IAM/Networking, the same as cloud practice • Developer UX (DX): key to widely spread • → Uniﬁed & integrated platform is required
policies ◦ → Use Terraform, Ansible, CFn, or whatever you want • Capacity management ◦ → “memory”-based optimization required; no practice (to me) • Chaos engineering framework ◦ → Use general chaos engineering tools