FR is truncating SQL even though I have disabled "SQL Text Limiting". What am I missing? (Also inconvenient in that you can't copy the sql text that is shown, you have to open "OpsPilot AI" to get the text.)

Curious, if you increase the 2000, does it affect the results?

No, I had it at 1000 and increased to 2000 and there was no change. When I copy what is shown I only get 1024 characters regardless of 1000 or 2000.

Where exactly are you seeing the 1024 limited text?

ahh, you're using the cloud UI

Yeah, well that gets me to my other issue I'm not able to easily find a given transaction in on-prem, because 1) I'm running 2 servers so I have to check both but more of the issue is 2) there's no way to filter for a given request, so I find myself having to look through a lot of requests. (I don't use FR much so maybe I'm doing it wrong.)

It depends on the amount of traffic on the server, yeah

But I do see the on-prem is not truncating so if I can find it then I should be able to get the sql I'm after.

on prem tracks all, but only keeps like 100 of the recent, slow, and longest transactions

Yeah, Cloud UI doesn't have much settings.

There may be some cloud settings in the on-prem UI. I can't recall.

on-prem Cloud Settings doesn't have much. Any nothing specific to sql truncation.

Well, how are you finding it in the cloud version?

Cloud UI shows "transactions names" and then I can drill into the one I'm interested in (and/or adjust the query params to get there). But on-prem, it's just a streaming list of them that disappear so hard to find one among a lot of requests.

Is the request in question neccessarily slow or erroring?

no

If so, there are dedicated screens for on prem to show only slow requests or only errored requests which don't "move as fast" as the standard history page

In any case, I did find the request by running it then quickly going to history. So I have what I need. Thanks!

FR does have an archived history of log files on disk which keeps every request ever, but it doesn't store all the details unforunatley.

FR keeps 100 transactions in the UI history, error, memory etc. lists - this is by design, because we're storing everything in memory. As @bdw429s rightly states, we also store these transactions into various log files, which are rotated every hour and we are stored (by default) for 30 days. In the cloud, you have more options. We're storing a sampled rate (default 95th Percentile) of all transactions. We are also capturing up to 10 so called interesting transactions each minute. These could be slow, or erroneous traces. Your best bet is to locate the transaction in the cloud if you're looking for something specific. I would also recommend looking at the Anomaly Detection capability - this will automatically highlight anything which is running a lot (Rate), causing (Errors) or running for a long time (Duration) - codename RED.

@davidtat Thanks for the reply, but the question was actually how to adjust the limit of the length of SQL displayed in the cloud

@davidtat To restate my problem, In FR Cloud, I can find a transaction easier because I can find it by "transaction name", but then I don't get the full SQL that I'm looking for. So then I need to use on-prem, which gives me full SQL, but I'm not able to easily find the transaction I'm looking for (especially if I have multiple servers with a decent amount of traffic).

@Nicholas Millard can you pick this one up Nicholas - I'm not sure it it's possible to adjust that setting for cloud - perhaps we have -D setting for it though...

Hi @philcruz you can use the JVM argument

-Dfr.observability.trace.value.limit=1024

(1024 is default, adjust as required) to adjust the SQL limit displayed for historic transactions saved in cloud. It will require a restart of the instance once added, and please be conscious that it may increase your trace data billing usage

@Nicholas Millard Thanks!