quantified.uncertainty/metaforecast
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

chore: add some more detail on different servers

Browse Source
This commit is contained in:
NunoSempere 2022-03-15 15:42:42 -04:00
parent 285c04f6ca
commit 1cba1ec35a
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
roadmap.md
View File

@ -28,14 +28,17 @@ Right now, I'm using various services to host various parts of metaforecast:
This could all be simplified. But this would require making an executive decision as to whether to go for scalability and greater expense, or to go for a cheaper but less scalable digitalocean server
I think that a simplication would be to:
1. Use one server for both external API and fetching probabilities each day
2. Use a Heroku server for the twitter bot.
2. Use a separate Heroku server for the twitter bot.
If I was doing this on my own, I would deploy 1. into a Njal.la server using Alpine Linux. However, more convenient options would be:
- A Digital Ocean droplet using Ubuntu/Debian
- A Heroku instance.
The biggest difference between Digital Ocean and Heroku is that I'm using [pm2](https://pm2.keymetrics.io/) on Digital Ocean, whereas Heroku has its own load balancer, logs, etc.
## General code review
Importance: 3/5, but good first issue. Would also allow you to incorporate