"508: Exhausted all retries for key: %s string" datastore error happens when I haven't gone over my limits

Ironically this just started happening for me as well, despite using alphabetic keys. It’s totally hit or miss but appears to happen more frequently in studio than in game.

1 Like

this kills my game

I’m getting this error too, absolutely no clue what is going on. Are you sure it’s fixed?

2 Likes

If this is still a problem, please file a Bug Report. Using this thread is not the appropriate answer to continue inspecting an issue that developers have no control over.

It seems to be a problem on Roblox’s end.

1 Like

When did you first receive this error? Does this apply to ordered data stores, standard data stores, or both?

Yesterday on Standard Data Stores. It’s intermittent, though, and changing key fixed it (for now). I can’t say whether it’s happening on Ordered Data Stores however, sorry.

4 Likes

This may or may not be helpful, but I am getting this error as well. I have two games using the exact same DataStore code and one gets this error when closing Play Solo in Studio and the other does not.

1 Like

I can confirm it’s happening for Ordered Data Stores.
Also this happened to me 6 days ago

2 Likes

Couple days ago, occurs on data stores, seems to be tied to the key itself, changing the key seems to work. Can read fine, issue is with saving.

Getting the issue with both OrderedDataStores and regular DataStores. It appears to be tied to specific keys. I posted a bug report with some info here: Certain DataStore keys repeatedly triggering "Exhausted all retries for key" error, unable to save

edit: merged below

1 Like

Earlier today, I noticed a massive spike of “Exhausted all retires for key” errors.

The error is 508: Exhausted all retries for key: %s HTTP 0 (HTTP 429 (HTTP/1.1 429 ))

Upon further investigation, I noticed that very few users were being affected by this very large number of errors. Then, I received this report from a player:

This player is repeatedly getting this error on one of their save files, and it appears to be specific to his key.

image

This error has appeared in both Vesteria and Miner’s Haven, the latter of which hasn’t had any update to its DataStore code for months.

3 Likes

This also happened to an old Data Store script of mines. I wonder if this could be the cause of the bug or is it specific keys or old keys would trigger this bug?

Happening to me at the moment, and I think its only effecting to specific keys.

Also worth noting: This error is not listed within the DataStore Errors page.

It’s hard for developers to handle errors when undocumented errors can occur. Of course, any developer should always be ready to handle unknown errors, but it would be nice if this was documented if it’s within the realm of possibility of happening again.

1 Like

Does this solve your problem? You’re unable to fix this problem by yourself as it’s on Roblox’s end.

It explains where the error comes from (internal) but doesn’t solve the issue that this error seems to be thrown a lot recently

1 Like

We looked into this and concluded there are two issues:

  1. Last week we introduced a bug where Data Stores started incorrectly returning 508 error codes no matter what the actual error was. We will be fixing this as soon as possible. This has no impact on the functional behavior of Data Stores.
  2. There has been a recent uptick in the amount of throttling for Ordered Data Stores. We have not made any changes to the system recently so we suspect that one or more games is causing disproportionate load on the system. We are investigating if this is the case and will be working with specific developers to understand their usage patterns. At the same time, we will be figuring out how to ensure that one game’s usage patterns do not disrupt other games.
12 Likes

Is there an estimated timeframe for the fix? testing using Play Solo in Studio is PAINFUL and SLOW.

Thank you!

For the first issue, we are planning to release a fix on servers sometime today. The change will not affect Studio until mid next week.

For the second issue, we do not have a concrete timeline yet. We are actively investigating how to proceed.

7 Likes