department-of-veterans-affairs / va.gov-team

Public resources for building on and in support of VA.gov. Visit complete Knowledge Hub:
https://depo-platform-documentation.scrollhelp.site/index.html
283 stars 204 forks source link

BGS profile accounts & payments #56199

Closed laineymajor closed 1 year ago

laineymajor commented 1 year ago

PROD

April 4, 2023 9:00AM ET

ROUTE

/profile/ch33_bank_accounts /profile/payment_history

MONITORING MECHANISMS

SUPPORT COMMUNICATIONS

April 4 at 9:00am ET, we will begin the VETS-API migration with EKS to PROD! We will only be rolling out with /v0/search.

Expectations of route-owning team(s): Report issues to #eks-migration-support Be available to answer questions we might have about your app and/or your reported issues(s) Be available to help us recreate an issue if needed, and test the resolution

Open Zoom Room! If at any point during the rollout, your team has an issue that needs in-person conversation, we will have a zoom room opn from 9:00AM-5:00PM ET. Don't hesitate to stop by and ask questions or talk through issues! April 4 zoom April 5 zoom

Monitoring by Tech Team 1

rmtolmach commented 1 year ago

👉 Mike C's Migration Dashboard aka "Chelen's Chart"

/profile/ch33_bank_accounts and profile/payment_history

Merged these two PRs at 10am ET:

And then kicked off revproxy Validated in datadog that the routes are being directed to the right place. /profile/ch33_bank_accounts and profile/payment_history 🎉 image image

That all looks good. No errors in ArgoCD prod.

IDENTITY users#show

PR merged:

and revproxy deployed. The logs are coming in and look good. Datadog logs image

With the addition of users#show endpoint, the portion of requests going through EKS has jumped up quite a bit! (it's was at about 1-3% before this.) Screenshot is from Mike Chelen's dasboard. image

sign_in and /sessions

Datadog dashboard Looks good: image

sidekiq

PR:

And then I turned the sidekiq workers down to 0 in AWS: image

Issues

no blocking issues, but a couple oddities seen in datadog

  1. One odd thing we saw was one weird blip where traffic should have 100% been going to EKS, but it was also going to BRD. The timing was at the exact time of the BRD and EKS deploy. There was also a scaling event at the same time. That blue part is BRD 😕 image
  2. The total traffic volume appeared to be lower for a period of a few hours today. It seems to recover at the end of the day. No one knows what's going on. It could be a metrics issue and not a real issue. image
laineymajor commented 1 year ago

ADDED TO THIS SPRINT AS PART OF THE EKS MIGRATION.