68 lines
3.1 KiB
Plaintext
68 lines
3.1 KiB
Plaintext
# Demo #0
|
|
|
|
The first demo of the full system in action.
|
|
|
|
## Part 1: Core Demo (Nov, 2016)
|
|
|
|
TODO: start with the "dumb" Docker Compose / Mongo demo.
|
|
- then add one by one to get to Mu
|
|
- persistence
|
|
- multi-node/sharding
|
|
- service discovery
|
|
- api contracts, pluggability, reuse
|
|
|
|
1. Muify an Express app plus MongoDB backend
|
|
Leverage Express and MongoDB containers
|
|
App already exists, just add a Mu.yaml (Compose-like) file
|
|
Show how adding AWS persistence to the MongoDB service is easy
|
|
TODO: starting point? Existing compose/cloud formation/kubernetes app?
|
|
TODO: talk about Mongo sharding, etc. as an implementation detail of mongodb/mongodb?
|
|
TODO: talking point on abstractions like key-value store (impl w/ Mongo, Dynamo, etc). Maybe show?
|
|
2. Import to my Marapongo account
|
|
$ mu remote add https://paxmu.com/joeduffy/demo1 && mu push
|
|
Shows up in UI sorta like a GitHub repo
|
|
I see a "deploy" button and a "live site" tab
|
|
3. Click the deploy button
|
|
Mention the capability for multiple environments
|
|
Deploys entirely to the cloud in <5 minutes (runs in AWS but customer needn't know)
|
|
To kill time, demonstrate running locally; talk to GCE, Kube etc.
|
|
4. Inspect the live site tab
|
|
Click production, see some simple "DataDog-like" analytics (requests per second, free disk space, etc)
|
|
5. Desire deeper analytics; add a "one liner" to my app to add ELK Mu stack
|
|
For bonus points, show searching MuHub for ELK in the UI
|
|
Click the deploy button, just deploys the delta (in <5 minutes)
|
|
To kill time, demonstrate that an AWS CF stack was created, click around CloudWatch, etc.
|
|
TODO: consume ES data via its API?
|
|
TODO: need more examples of the daemon concept... less infra, more app-oriented.
|
|
6. Reinspect the live site tab
|
|
Now I see lots of great Elasticsearch data (IFRAME-like), etc.
|
|
|
|
## Part 2: GitHub Integration (TBD)
|
|
|
|
1. Hook Marapongo up to my GitHub account, so I can use GitFlow for deploys
|
|
Use a GitHub "workspace" (monorepo) to contain multiple services
|
|
Link branches to environments
|
|
Simply push a branch to trigger CI/CD workflow
|
|
GitHub protected branches and automatic propagation from stage to prod
|
|
|
|
## Part 3: Serverless (TBD)
|
|
|
|
1. Now we want to set up alerts; navigate to the "alerts" tab
|
|
Ability to set up "PagerDuty-like" alerts
|
|
Log an issue and send message to Slack channel when condition occurs
|
|
Lambdas/Triggers from the ELK stack (e.g., https://github.com/Yelp/elastalert)
|
|
Demos the Riff-like serverless capabilities of the platform, incl SaaS integration w/ Slack
|
|
|
|
## Part 4: DSL
|
|
|
|
1. Refactor the whole thing to have zero metadata, just beautiful code
|
|
|
|
## Part 5: Sharing and Reuse (TBD)
|
|
|
|
1. Share my template with my other teammates
|
|
TODO: carve up in half: 1/2 reusable boilerplate, 1/2 app logic (routes.js-like)?
|
|
Refactor to be generic "Express, MongoDB, and ELK" template, publish in Marapongo
|
|
Show how cloning and deriving a new project off it is as easy as "base: myteam-jselk-stack"
|
|
Now also show sharing and reuse of our "PagerDuty stack" just like others ($ mu add pagers)
|
|
|