very frequent SSL handshake failure is rendering dynalist web service unavailable
I lost a lot of important work notes several times in the past 1 month
Hmm Iāve been battling with these issues lately trying to get it to stabilize. If you are working and this happens, donāt close the tab or refresh - keep the page open and it should keep working (despite not syncing) and it will sync as soon as service is restored.
Hi
Is the issue resolved?
I would love to come back to using dynalist as my daily driver, if this is resolved or going to be resolved soon
It still happens from time to time and we are still trying to figure out if itās a hardware issue or networking issue that the server is losing database connections. But we have monitoring in place and means to reconnect if I am close to my desk.
For now your best bet is to keep the page open, wait for it to sync, donāt refresh it. Or use the apps, which I know have some long running bugs that we havenāt had a chance to look at yet, but at least it works offline.
Thanks for looking into this and also for suggesting the work arounds for dealing with the issue
Hope to see this resolved soon
This is getting worse. Used to be Iād see it every 6 months. Now 1-4x in a 2 week period.
I have spent a lot of time recently trying to understand the recent outages. Unfortunately, this seems to be due to two things:
- A few users seems to be making heavy use of the API to repeatedly refresh their data, which causes heavy reads to the database.
- And a few users who have a lot of documents and also shares a lot of documents/folders. Due to some unfortunate internal designs done long ago, which are not very easy to change now, these users incur heavy loads to the server in bursts when they make changes to their folder structure.
We will be adjusting API rate limits as well as reaching out to a few heavy users to help them find a better way.
1 Like
@Shida glad to hear that, thanks
Hereās a 520 from today; not something Iām used to seeing, from Dynalist or anywhere else.
Sounds like the throttling could help a lot. IDK if thereās enough such users that it makes sense to have another pricing tier for them.
Edit: Fortunately, looks like this went away after just a few minutes.
Yeah the throttling is definitely helping, but unfortunately some of these users have enormous documents and workspaces that just loading it could set the server back by tens of seconds. So far, the throttling is doing a good job at preventing these single users from bringing the server down, but sometimes it still happens where multiple users with extremely large documents happen to load around the same time (for example, during the start of a work day).
Weāve started reaching out to some of these users to request them to tone down the usage (by deleting no-longer-used data, un-sharing any shared documents, and breaking apart large documents), so it should get better over time.
1 Like