[Deployment Infra] Continuous Deployment of PATH

Olshansk commented 2 days ago

Objective

CD of the latest HEAD on the main branch.

Origin Document

TODO: @okdas to link to documentation related to staging deployment of poktroll

Goals

Continuous deployment of PATH
Ability to manually test the latest deployed version of PATH
Allowing anyone on the team to maintain this

Deliverables

[ ] Deploy the latest build image on main of path to a staging environment
[ ] A README explaining the end-to-end flow, how to track it, identify issues with the infra, update it, etc...
[ ] Leverage the same end-to-end tooling as in poktroll
[ ] A make target that sends a request through the staging instance of PATH.

Non-goals / Non-deliverables

Load testing the infra
Continuously stress testing the infra

General deliverables

[ ] Comments: Add/update TODOs and comments alongside the source code so it is easier to follow.
[ ] Testing: Add new tests (unit and/or E2E) to the test suite.
[ ] Makefile: Add new targets to the Makefile to make the new functionality easier to use.
[ ] Documentation: Update architectural or development READMEs; use mermaid diagrams where appropriate.

Creator: @Olshansk Co-Owners: @HebertCL @okdas

HebertCL commented 2 days ago

Briefly describing an action plan for this issue:

Create a pipeline that ensures a staging deployment of path will take always the latest image (planning on using ArgoCD image auto-updater)
Add a K6 tests script to be executed against the freshly deployed image. Depending on the time available, we can either make this a co-located script within the repo we execute on demand, or plan to make this as some form of integration testing phase properly wired to a pipeline.
Document the whole setup for both deliverables previously described.

Some of the questions that come to mind before actually jumping into this:

Are we planing on using the already existing portal infra? Do we want to spin up new infra for path?
Are we planning to add this to our existing private Grafana? Do we want to spin up a new one for this?

Olshansk commented 2 days ago

Create a pipeline that ensures...

👍

Add a K6 tests script to be executed ...

Let's not overcomplicate and use something in the Makefile or add a new one if it doesn't exist.

Document the whole setup for both deliverables previously described.

👍

Are we planing on using the already existing portal infra? Do we want to spin up new infra for path?

Biasing towards new infra.

Can you list out (in < 4 bullet points), what the tradeoffs are?

Are we planning to add this to our existing private Grafana? Do we want to spin up a new one for this?

New grafana

HebertCL commented 1 day ago

Are we planing on using the already existing portal infra? Do we want to spin up new infra for path?

Biasing towards new infra.

Can you list out (in < 4 bullet points), what the tradeoffs are?

I am also biased towards new infra. Some reasons for it:

Old staging infra hosts portal-middleware. So we will be mixing the old-private codebase with new one.
Recycling the old infra might lead us to carry over any pitfalls and issues these setup has (either identified or not yet identified.)

buildwithgrove / path