Good call. Will include in next Annoyingness Straw Poll!
As for goal tagging, this is not a complete solution but we do have a kind of way to do some tagging:
Part of the problem is that I’m not even sure all of the confusing ways this can break. Possible insta-derails if data has been edited, that sort of thing. Also I think a blatant bug with choosing the new pledge amount when restarting. This is actually high priority regardless of what the poll says, since it has a particularly high cockroach factor.
That does sound bad. Just wondering, why are insta-derails even a thing? I’m pretty sure I’ve never had one that I didn’t email support to undo. I’m having trouble imagining the case where a goal should derail at a time other than (the next run of a batch job after) its deadline. At least until we have an uncle button.
I would also love it if derails were limited to the update time; it would give me a chance to fix failed road edits, for instance, or play around with a restarted goal without screwing myself over. Or even change data without accidentally derailing myself!
Seriously @dreev, this has gotten worse and worse now. I don’t think there is any time of the day that graphs are produced in a logical amount of time. It often takes more than 2 minutes! Time for a server upgrade?
@bee’s been hard at work on tracking down the slow graph regenerations. We were on a bit of a wild goose chase thinking it had to do with our queuing implementation for when we’re processing lots of graphs at once. Turns out our database server was just on its knees frequently. We upgraded Mongo today and (fingers crossed) things look a lot better. Like 5 seconds to regenerate a graph, which is still a stupidly long time but we’re hopeful that we’ve at least solved the problem of frequently minute-long refresh times. Keep letting us know how things look from your perspective!
Well, queueing and database stuff is critical regardless. Maybe it would be moot with client-side graphing taking a ton of load off the server but the proof-of-concept we made had some fatal flaws (plus we need the server-generated version as well for various things like the phone apps, hotlinks, etc) so we’re now taking our time and doing it right. So for now we’re making sure the server-generated graphs at least take a non-painful amount of time, ie, a few seconds.
I hope you implement “making it easy to change settings for multiple goals at once” sooner rather than later. I think it’s holding me back from reviving my goals on Beeminder (and thus lowering your potential revenue), as when I think of having to edit a lot of goals when I go on vacation, I think “Ugh”. Not a good sentiment to associate with a product.
