I got the AWS bill for July, and it’s not quite as amusing as the one for June was. It’s not going to destroy the project, but it does show that I need to keep an eye on costs.
Going through the significant line items, CloudWatch is one that I didn’t really want to see there. CloudWatch is cool – it’s where a lot of the graphs that I post come from. I had 4 dashboards showing me graphs, and it seems I only get 3 for free, and the 4th cost me $2.87. As I didn’t use it anyway, I trashed it.
The other component of CloudWatch is a charge for data sent to the lambda logs. It’s not for storing that data, it’s just for getting it there. I was warned weeks ago that that was going to be a problem so I cut down the amount of logging I was doing. Still, there has been a lot of activity this month due to the plays download, so there was always a bit of logging from that. Let’s see how it goes next month.
The Data Transfer costs are small so far, but I’m not sure whether they’re for me downloading data or for me sending out requested data. If it’s the latter, the site is in trouble when I start advertising it to 3000 users. Let’s watch that one.
The Elastic Compute Cloud costs are for the blog. I realised that cost was coming so during the month I decreased the size of the virtual machine that the blog was hosted on, and that cost should be less next month.
The costs for Lambda are for way exceeding the free quota of time and compute power. I suspect that is also related to the plays download.
The Relational Database Service costs are for the database, and that one will probably go higher next month. As mentioned in previous posts I had to upgrade the database during the month so that it could cope with the plays download load, and it’s still not doing that well. The good news is that for about double the cost I think I can start using AWS Aurora, which allegedly has very good performance. So when I organise a revenue stream for the site that might be resolved. Until then it will be coming out of my pocket.
As mentioned last month, it’s not too bad, but the costs definitely remain one of my priorities. When I start asking for Patreon contributions and putting ads on the site, you’ll know why. It’s not because I’m trying to profit from the site, but because something of this scale costs money under any circumstances.
I know this has been said before, but I’m totally ready to pitch something toward crowd funding the site when you get to it.
Thank you Michael. For others reading this comment, there’s a Patreon link in the navigation bar on the site.