We launched PasteThePlan.com on September 11th, and it got just shy of 5K pageviews in 3 weeks:
After launch, it stabilized at around 125 pageviews per weekday. (That’ll bump up this week with a new page that lets you see the 20 most recently pasted plans. I kinda find that addicting.)
In Richie’s architecture explanation, he wrote that PasteThePlan runs on Amazon Web Services:
- The web requests are handled by AWS Lambda, aka serverless code or function-as-a-service
- The plans are logged in AWS DynamoDB, a managed NoSQL database
So what’s it all cost to host?
(Note: two very tiny costs aren’t included. We’re storing the plans in AWS S3 file storage, and backing them up in Azure File Storage. These cost pennies per month. At some point, I’ll need to break those out separately on our cloud bills, but right now the costs are so doggone low that it doesn’t even make sense to tag those resources for billing purposes.)
When we launched, commenters asked a lot of questions about “why didn’t you use SQL Server?” I think this pretty well speaks for itself, and it’s what I meant when I wrote that our community projects have to be sustainable. This project is completely sustainable, and will never need fees, ads, or sponsors.
“Why didn’t you use SQL Server?” a: “because I do not CRAP MONEY. thank you.”
You mentioned recently (a month or three ago? can’t remember exactly where) that your vision was for BOU to be a part of every day of every MSSQL DBA on the planet. (Or something like that). I imagine that this service can lead to things like:
an “analyze” button that runs lots of checks (kinda like an “sp_BlitzPlan”) on the submitted plan looking for all sorts of anti-patterns? e.g. “Your actual plan’s actual/estimate diverges sharply at this join… have you tried creating a multi-column stats object on those cols?” Maybe even a link to a great technical post for each issue found (again very sp_Blitz-ish).
once that’s in place, a “most common issues found” report that aggregates across your submitted plans, listing great technical resources for those issues and essentially creating a “query plan training program” for those looking to get better at QPs?
If so, that sounds like a ton of fun… something I don’t think’s been done before.
Bingo! And that’s about all I’ll say on that one. 😉
[…] generally cheaper and easier to manage than relational databases. Yeah, I said it. Come at me. Try hosting production-quality, auto-scaling relational databases for $3.43 per month. It’s damn […]