Confirmed by root cause analysis in Attic updates fail when using 2026-06-06 database clone · Issue #792 · drolbr/Overpass-API · GitHub. The 2026-06-06 clone had attic data that was not in sync with the rest of the database and that caused the crash.
2 Likes
Hi @Kai_Johnson
I deployed it on my local machine, as a test, and it was much faster, and barely loading the machine at all. So, I then deployed it to an EC2 (t3.large) with GP3 storage, and it is also much faster than my ECS+EFS was. Maybe approx 30x. And increasing the throughput did not significantly increase response times, like I saw on ECS+EFS.
It will cost me more to run it, but not much.
Thanks for your help - Brock
1 Like
That’s why it’s often easier to just regularly do a full re-import of the overpass database (e.g. weekly) which is usually what I do.
Yeah, that was the pain point that prompted me to put all the work into the improvements in the container build. It’s much more stable now. I can run it pretty much indefinitely without issues as long as some external event (like a power outage) doesn’t disrupt it.
If you get tired of those weekly re-imports, you should try the container sometime.
There are some significant improvements in the latest container image out today. Most instances should see some modest performance improvement from the latest optimizations. And there are some powerful new monitoring tools if you’re willing to do a little setup to plug them in.
2 Likes