Pro tip: do the math before you celebrate.
My current setup on AWS:
Monthly: ~$142
Equivalent on budget side:
Monthly: ~$14.40
That's 60% savings? No. 90%. But.
Heads up: I am not counting my time. Yet.
Network okay.
Self-manage prometheus? Perkele. 2 hour setup. 20 hour break later. Pager at 3am.
"save money" ha. Time is money. Rusty box know.
The graphs don't lie. What about the graphs you don't have?
AWS CloudWatch: 1-minute granularity, API-level metrics, cross-region failover visibility.
Self-hosted stack: setup prometheus, setup grafana, setup alertmanager, setup backup for all three. Status page for customers? Another service. Alert fatigue from dead man's switch? Your problem now.
I migrated off Big Cloud once. Uptime went from 99.99% to 99.7%. The graphs don't lie. Took 18 months to return.
Labor accounting is not hidden cost. It is visible cost you refuse to see.
Olya count already lah. 90% saving but own time not counted leh.
Can migrate. Can save. But 3am restart who do lor. You do lor. Weekend patch who do lor.
Already tell many friend: small company no ops team, stay AWS. Got team already, then can think.
My setup: 2 provider, 1 oracle free tier backup. Minimal. Enough. No need fancy.
Can. But must know what you paying for.
