overview for gunaa

Running a full SaaS stack for $3.32/month exact breakdown by gunaa_dev in c/opensource@programming.dev

[-] gunaa_dev@programming.dev 1 points 12 hours ago

zewm @lemmy.world

Account age 1 day. The entire post is a sales pitch. Go eat shit.

sorry did not mean to offend anyone if it's not the right place for it i get it

Portabase now handles backup/restore of Docker volumes by gunaa_dev in c/opensource@programming.dev

[-] gunaa_dev@programming.dev 2 points 18 hours ago

ahhh for the server-based ones for sqlite specifically i'm just doing periodic copies of the db file to supabase storage as a backup, nothing fancy

Portabase now handles backup/restore of Docker volumes by gunaa_dev in c/opensource@programming.dev

[-] gunaa_dev@programming.dev 2 points 21 hours ago

Nice timing I'm running SQLite + Redis on Railway and was just thinking about backup strategy. Does Portabase handle Railway volumes or only self-hosted Docker environments?

Running a full SaaS stack for $3.32/month exact breakdown by gunaa_dev in c/opensource@programming.dev

[-] gunaa_dev@programming.dev 1 points 22 hours ago* (last edited 21 hours ago)

onlinepersona

Would be cool to add peertube, loops, and pixelfed as a destinations.

opened a github issue to track all three. pixelfed is going in first since it uses the mastodon api so most of the work is already done. peertube after that. loops

-16

Running a full SaaS stack for $3.32/month exact breakdown (github.com)

submitted 1 day ago* (last edited 21 hours ago) by gunaa_dev@programming.dev to c/opensource@programming.dev

4 comments fedilink

so i've been running posthive + 2 other projects on railway hobby plan for a while now and was curious what it actually costs

here's the breakdown:

posthive api (fastify + bullmq + sqlite) sitting at 136mb ram-> $0.59/month redis for the job queue, only using 12mb memory, 300mb volume for rdb snapshots -> $0.12/month two other projects -> $1.63/month

total so far this month: $2.58, estimated end of month: $3.32

the thing i like about railway billing is it's based on memory + cpu time, not requests. so traffic spikes don't really affect the cost. egress is free tier too.

kept redis lean by configuring bullmq to only keep the last 100 completed and 200 failed jobs so it doesn't just accumulate forever

for the docker setup i'm building multi-arch images (amd64 + arm64) on ghcr with github actions layer caching, and a paths filter so doc-only commits don't trigger a rebuild

source: github.com/AstaBlackClove/posthive