As written in Sentry Pricing,
hobbyist’s History is 7days,
and small and the others’ history is 90days.
But from the start, what’s the ‘history’?
Is this mean error’s lifespan?
For example, I pay for the small plan, and an error comes on 4/1/2018,
so the error will disappear on 7/2/2018?
Can someone please confirm what history means? I also have the impression it implies that errors will disappear after 7 (hobby) to 90 (all other plans) days, but I can’t really imagine that anyone really wants this…
Its retention – it means you don’t have access to data older than 7 days. What this actually translates to in the UI is a best less obvious given we aggregate data.
For example, if an error happens 9 days ago, and then happens again 6 days ago, you’ll still know that it happened 9 days ago, you just wont have a sample from that time period. On the other hand, if it happened 9 days ago and hasn’t happened since, you will not know it ever happened.
Thanks for the quick reply and clear answer! Being on the “hobby” plan I know I can’t ask for anything but just to provide my feedback: I’d much rather have full history with only a tenth of the errors/ month than the current structure. Also, if I move to a paid package, I’d expect errors to persist forever (as awful as that sounds ;)).
We primarily think of Sentry as a paid service which is why we do this. It’s actually quite expensive to retain and process errors given the large payload size and some of the preprocessing that has to happen. The hobbyist plan to some degree is our offering based on principals: we want everyone to use Sentry, thus we keep a hugely generous free plan. The restrictive features and retention are our version of saying “please pay us if you find value in Sentry for your team” 
(A long time ago Sentry used to be paid-only unless you ran the Sentry service on your own hardware).
Hi - we just had an alert for an issue that I know I’ve researched before. So I looked and low and behold I couldn’t find it. Looked in our slack Logs, and it seems that this previously happened back in May this year.
So - thanks to our site having a low level of issues I happened to remember that I had previously researched and resolve / ignored the error (it seems to be caused by a browser extension) - but this new error doesn’t seem to tie to that previous error.
This seems contrary to you saying that I would know it had happened previously - has something changed since April?
thanks
I get your point, but it still doesn’t change the fact that it’s a free service. If you need longer retention I’d suggest upgrading to a paid tier or running Sentry on your own servers. We do, after all, have our own server bills, engineering time, etc that we need to pay for.