2024-01-31: office day

  • add GHA to ce-nonprod: https://github.com/elsevier-centraltechnology/tio-terraformcontrol-ce/pull/1212/
  • appraisal
  • K8s 1.26 support (Mars PPE still flagging issue after release)
  • okr ideas
    • integrate advisor & crtxctl, benefit: jump in visibility / usefulness of Advisor
    • deployment pipeline pattern
      • prepare (terraform), deploy (helm), verify (capability tests), report (advisor? operations website?)
    • cortex bpm

2024-01-30

  • CEIP-4469: KSI migration
    • brainstorming about how to roll KSI out on Kong
    • generalise GHA for Helm further
  • terraform for GHA on ce-nonprod (for flowable)

2024-01-29

  • CEIP-5073: registries check
  • CEIP-4469: KSI migration
    • tidy up GHA for newrelic
    • move on to GHA for Kong, then return to clean up terraform (if GHA approach approved)
    • GHA
      • take as input the datestamp to replacei, ie treat blue-green as infra only?
  • Retro

2024-01-26

  • CEIP-4469: KSI migration
    • PoC GHA helm deployment
      • role-to-assume: arn:aws:iam::595468393306:role/Core-Elsevier-Platform-Manager-Role-nonprod Error: Could not assume role with OIDC: Not authorized to perform sts:AssumeRoleWithWebIdentity Solution: https://github.com/elsevier-centraltechnology/tio-terraformcontrol-ce/blob/gha-helm/595468393306/oidc-github-actions/kong-image-build-iam.tf Specifically:
        "Principal" : {
    "Federated" : aws_iam_openid_connect_provider.github_actions.arn
  },
        GHA working but with a couple of apparently ignorable errors!
        E0126 17:16:35.028853     348 memcache.go:287] couldn't get resource list for projectcalico.org/v3: Unauthorized
E0126 17:16:35.029442     348 memcache.go:287] couldn't get resource list for metrics.k8s.io/v1beta1: Unauthorized

2024-01-25

  • CEIP-4469: KSI migration
    • successfully completed kong_db_yyyymmdd
    • moved onto control plane
    • 74 min terraform: run out of security groups?
  • CEIP-5105: move to Artifactory tokens
    • created token and proof of concept
  • GHA helm deployment
    • cortex-prod-admrole
    • kong-nonprod-beta
    • helm install

2024-01-24

  • CEIP-4469: KSI migration
    • completed the jenkins ksi removal including removing two man, seven year old container image
    • start same approach on kong-infra
      • kubectl get job postgres-20221121-init -n infra -o json | tfk8s --strip -o postgres-20221121-init.tf

2024-01-23

  • CEIP-4469: KSI migration
    • Solved CSI terraform issue for newrelic:
    • Consolidated newrelic-logging
    • start on jenkins-infra
      • discovered cannot inject volumes into init container of jenkins chart (direct never mind via fulcrum)
      • onion problem, start peeling…
      • considered bringing the jcasc checkout into jenkins pipeline but don’t understand enough about it (why is default config disabled, hence cannot start without an initial checkout)
      • ended up reading secret within init container
      • TIL: faster to dev by using helm -f values then convert back to terraform

2024-01-22

2024-01-19

  • CEIP-4469: KSI migration
    • kong new relic: stuck on terraform error: Error: YAML parse error on nri-bundle/charts/newrelic-infrastructure/templates/kubelet/daemonset.yaml: error converting YAML to JSON: yaml: line 68: did not find expected '-' indicator
    • KT with jonathan

2024-01-18

  • CEIP-4469: KSI migration
    • grafana: TIL: kubectl get secretproviderclass backstage-dev-secret-provider -n backstage -o yaml | tfk8s --strip -o backstage-dev-secret-provider.tf

2024-01-17

  • CEIP-4469: KSI migration
    • a lot wasted on terraform debugging LOATHE: terraform state allows working code on one platform (darwin_amd64) to be unsupported on another (darwin_arm64) LOATHE: this is made worse by much of our terraform being ‘abandonware’ with no ownership or maintenance LOATHE: terraform, at least as done at ELS, is fundamentally broken:
      • code is deployed before being reviewed or committed.
      • modules are arbitrarily sized and structured
      • testing does not exist

2024-01-16

  • CEIP-4469: KSI migration
    • working on various backstage projects

2024-01-15

2024-01-12: vacation

2024-01-10,11

2024-01-09

2024-01-08

  • new machine setup
  • CEIP-5002: exclude FlowSchema items from 1.26 migration report

2024-01-05

  • CEIP-4979: crtxctl generic JSON renderer: apply to all existing commands
  • new machine setup
  • Camunda:
    • attendees
      • Graeme Wilkinson (commercial, Camunda)
      • Anton von Weltzein: (sales eng, Camunda)
      • Laksmi Remani (sales eng, Camunda)
      • Richard
      • Sravan
      • Rakesh
      • Prasath
      • Tim
    • agenda: product and value prop.
    • q: why no business users?
    • open stds, faster time to market, retain existing tool choices
    • modeler:
      • organised into projects
      • collaboration tool
      • analogy to google docs
      • automaatic change history (again like google docs)
      • process, decisions and forms all together
    • Richard questions
      • due date of task?
      • conditions based on what?
      • scoping of data?
    • SaaS version must use Optimise API to extract (or capabilties built into UI)
    • gRPC gateway
    • connector sdk has some capability to inject data and obviously process data goes to elastic search
    • camunda 7 optimise has an advantage here in being able to ingest customer data

2024-01-04

  • run 1.26 migration report
  • CEIP-4979: crtxctl generic JSON renderer: single implementation

2024-01-03: Vacation