2022-08-31

  • Kong update w Felipe

    • hard limit of 120 rules per ASG

      • if had done portal would have seen already so need to fix

    • cluster and cluster telemetry need load balancer

    • proxy nice to have LB

    • what about classic LB (sim. to ALB)

    • all now working

    • infra running ok

    • moving sandbox to blue, then nonprod, prod

    • then I am to commission RDS and the rest follows

    • should have had dev portal by w/c 29/08

    • Health

      • had 4 sets of APIs in one environment
      • deck removing one as deployed the next -> deck-sanity

    • moving from blue to green will should be just deck dump, deck apply

  • merge FG PR

  • resolve remaining conflicts

  • look at validation

    • what does workspaces.yaml dp_fqdn mean?
    • nothing.
    • invert? what data plane targetting from workspace
    • A workspace publishes APIs in 0..n data planes

  • commission RDS

  • draft PR per environment for new workspaces.yaml and get everyone to approve

  • C3

    • Platman 1.0
    • Garrett promises slides to follow
    • no more pause reconciler
      • no more continuous reconciliation because don’t know state
    • move to event based reconciler, two triggers
      • push by team
      • push by CE (notified)
    • managed node groups
      • asg groups managed by AWS

2022-08-19

  • CEIP-2173
    • GitHub actions implementation done and merged
  • The Book

2022-08-18

  • dev duty
  • CEIP-2176:workspaces.yaml validation
    • python3 -c 'import json, sys, yaml; from jsonschema import validate; validate(instance=yaml.safe_load(open("workspaces.yaml")),schema=json.load(open("workspaces.schema.json")))'
    • first with Tekton
    • then with GitHub actions
  • Conversation with Len
    • Prod Mgmt
      • James Coates (DBS) have their own TIO, more directed towards Cortex
      • The other BUs have cortex mostly tio-led and hence less keen or able to influence
        • Tio cross-bills for DevOps so a good pitch for cortex is to save budget for prod dev’t
    • Speak to Irfan about ‘the book’

2022-08-17

  • CEIP-2175: contacts teams and writing blog post

2022-08-16

  • many meetings about whether to migrate or recreate clean Kong database

2022-08-10

  • CEIP-2120: delete workspace
    • new from Kong support: kong migrations reinitialize-workspace-entity-counters
      + /docker-entrypoint.sh kong migrations reinitialize-workspace-entity-counters
Error: admin_gui_session_conf must be set when admin_gui_auth is enabled
      I think the only way is to do this in postgres: Take a backup
      select id from workspaces where name = 'here';
update workspace_entity_counters set count = 0 where
workspace_id = 'id_here';

2022-08-09

2022-08-08

  • AM: CEIP-2120: working w JB on the deletion of workspaces
  • PM: CEIP-1995: working w AN on the resultant Jenkins agent crash

2022-08-03

  • CEIP-2130
    • what about a k8s crontab?

2022-08-03

  • Account compliance dashboard mtg:
  • CEIP-2130
    • investigated GitHub Apps example
      • local server doesn’t start on Linux
        mnt/data/git/GitHub-app-webhook-monitor/vendor/bundle/ruby/3.0.0/gems/mustermann-1.0.3/lib/mustermann/regular.rb:22:in `initialize': wrong number of arguments (given 2, expected 1) (ArgumentError)

from /mnt/data/git/GitHub-app-webhook-monitor/vendor/bundle/ruby/3.0.0/gems/mustermann-1.0.3/lib/mustermann/pattern.rb:59:in new' ``` - smee doesn't work with ZScaler (have to exit not merely turn off) - ruby 2.6 gives this error: ``` /Users/stephensont/git/GitHub-app-template/vendor/bundle/ruby/2.6.0/gems/octokit-4.13.0/lib/octokit/middleware/follow_redirects.rb:14:in module:Middleware’: uninitialized constant Faraday::Error::ClientError (NameError) Did you mean? Faraday::ClientError ``` and brew installing Ruby 3.1 fails as on Linux - CONCLUSION: GitHub apps are more trouble than they’re worth

  • CEIP-1995
    • how far has rollout gone?
    • JVM parameters (Jenkins on Java 11)

2022-08-02

  • CEIP-2130 Jenkins envelope ok, I can build a GitHub App to monitor failures and do something (retry on a fibonacci delay and / escalate on slack) or argue the case with Matteo or fork the GHF to queue the requests ?
  • CEIP-1995 and CEIP-2137: Jenkins outage and memory resource changes

2022-08-01

  • dev duty
  • ticket for Jenkins

2022-07-26