Production Venue Deployments

[mike-gangl]

Edit: Manifest of Deployed Capabilities across all venues

Venue Services

• Each venue needs an airflow instance (manually or via marketplace), and that needs to be integrated with proxies (API Gateway, HTTPD)
• Airflow needs to be integrated with cognito, not the custom username/password it's using now.
• Each venue needs a jupyterhub (manually or via marketplace), and that needs to be integrated with proxies (API Gateway, HTTPD)
• Each venue needs a navbar/healthcheck (manually or via marketplace), and that needs to be integrated with proxies (API Gateway, HTTPD)
  • All services need to be integrated with proxies (/unity/dev, unity/test, unity/prod)
  • All services need to have SSL enabled on their load balancer(s). This is available in the unity-venue-dev account right now.
  • we will manually add proxy information to the venue proxies until UCS has an automated way of doing this.

Shared Services

• U-DS available and at latest versions
• App-pack-gen endpoints
• Dockstore (TBD)

Infrastructure (venue infrastructure via management console)

• Cognito user pool
• HTTPD shared service proxy (dev, test, prod) and 'www' DNS entries created for them
  • Proxy info is manually updated and edited for now.
  • configured with mod-oauth
• HTTPD venue proxies deployed via management console
• HTTPD Venue proxies integrated with shared service proxies (manual)
• Venue API Gateways deployed via management console
• Venue API Gateway integrated with shared API Gateway (manual)
• Instructions on adding routes/proxy info to API Gateway and HTTPD

Need PRODUCTION VENUES for EMIT, SITC

At PI: Need a shared timeline to which we are accountable for this effort. SPIKE this - essentially all teams should be focusing on this effort at 100%

[mike-gangl]

• Once we do deployments into dev and test, all deployments to operation venues should be measured for the time it takes to stand up and integrate venue features. Essentially, each team needs to measure the time the venue is available (to them) to when their services are accessible.
• Multiple (2) teams being supported in the MDPS operational environment.
• For MCP tickets coming through that impact production venues, we shall note the time of ticket creation, time of ticket closing (on our end), and time of ticket closing on MCP side.
• Weekly spend reports by service areas