I have a bad news.
Again I get hundreds of reports from players that they cannot play due to CloudScriptAPIRequestErrors.
This time they are getting "Temporary storage issue, please retry request» errors.This is the second fail in November.
Some players, including payers, swear, demand compensation and leave the game.
Since November, I have to pay almost $ 2,000 instead of 0. But the service has become terrible!
Could you please clarify which Title ID this is for, and provide more details on the calls you are making and the results? The issue you're referring to concerning the Cloud Script error from your previous post was fixed that same day, and has not re-occurred since. Looking at your "Raid Heroes: Total War" game, I do see excessive calls to Entity updates in the title, which is producing some errors - you can see this in your Event History by searching on title_exceeded_limit. But those errors go back more than a month, so that would not be new behavior.
As to your payments, where are you seeing that, exactly? As you pointed out, you were in the legacy free tier until this month, so you have no payments on file, and your current billing summary page shows only minor overages - largely on the Profile Write meter (though based on the usage pattern in your titles, I would expect that to increase significantly before the end of the month, and I would highly recommend optimizing the titles to reduce overall usage).
Hello, Brendan!
Thanks for the reply!
As I understood from the answer to the support ticket the issue from my previous post was not fixed. You have just rolled the update back. I was asked to follow up on problems for a week. And there have been repeated peaks in runtime and player reports these days.
But yesterday the error was not the same as in the previous post. I received almost 500 reports from players with error text: "Temporary storage issue, please retry request». These are only those players who themselves sent the report to me by mail. There were much more errors.
Titlde ID D984B
I see CloudScriptAPIRequestErrors peaks in the dashboard:
I'm really not sure how to interpret your first statement. When we find that an update introduced an issue, the fix is specifically to roll back that update, so that the error is no longer present. We then investigate all the code changes in the update to determine what the specific cause was and how it wasn't caught by the existing test suite, so that we can add tests to prevent it in future. Is there some other way you would interpret "fix" other than "take action to stop the error from occurring"?
Thanks for the added details on this issue. You've got two other errors listed above:
1. For the temporary storage issue, that's directly related to the error I mentioned above - the too-high rate of attempted updates to the Entity data for the player accounts. Again, you can see this by searching on title_exceeded_limit in your event history. Fixing those would fix the data error.
2. For the Cloud Script error in the graph, are you logging the specific error from the script using a try/catch? Since it's "updatePlayerData", my suspicion is that it is again related to the updated (see 1), but we'd need more info to be sure.
The second point is indeed related to the first. But my code worked for almost a year before and I never got "Temporary storage..." error. The graph that I attached above shows that the error peaks are of a one-time nature. I have never seen them before and today they were not. Since I have not made any updates on my part, the failure was probably on your side. Also reports were sent from hundreds of different players. Therefore, I also exclude the possibility that someone was doing something non-standard in the game.
Now my game is sending a request to save data every 100 seconds (if there is something to save). UpdateUserData, UpdatePlayerStatistics and UpdateObject calls are used (one per save attempt). Additionally, there may be other calls related to saving battle results. But according to statistics, there are hundreds of times less of them than save calls. Therefore, they cannot lead to problems.
Moreover, until November, the game sent save requests every 2-3 seconds and I did not observe any problems for 2 years of playfab usage. All these problems started in November. Therefore, now that the service has become paid for me, I feel indignation. Sorry for the harshness.
None of the limits have changed (we have a strict policy against breaking changes, and that would be a breaking change), so we should work through the logic to see why you're seeing something different.
To start, are you saying that you never had title_exceeded_limit errors before? If so, that would mean something in the client logic changed that increased its call rate. Is that rate dependent upon anything, like how many games a player played or how many other players they've played with? We can look at older logs to see if the title_exceeded_limit goes back beyond the 30 days you can see.
If you're not sure about title_exceeded_limit, do you mean it's just that the errors you're now seeing weren't there before (or possible weren't noticed due to there being few of them)? My belief is that this is the case, since at very low volumes, data update rate limit errors will be limited to the event and may not show up in the API graph. It's only at higher volumes that you'd start to see the knock-on effects, like database throttling.
I had exceeded the limits 1-2 times a month earlier (when I updated the player data every 1-2 seconds). But there have never been massive reports from players with the "Temporary storage ..." error.
At the end of October, I optimized the number of savings to 1 per 100 seconds. This reduced the total daily number of API calls by 4 times. This problem appeared about 10 days after optimization
3 People are following this question.
