previously I was able to use regex for extracting parts of URL and replace them with some identificator. I have to DT clusters on the same version 1.176. One of it has old settings with regexes, one has new one when I can define placeholder with extraction or replace rules. But “between” is not always ok, because sometimes it needs to prepare multiple rules (hundreds of uris, that are hard to cover with single url, we have multiple UUIDS there). I don’t see updated docs about this topic.
Julius is right. We have no plan to bring regex back, but we are locking for improvements on the new naming patterns. @sebastian k. the regex you posted is that the one that you need to recreated?
Yes, without this I have to create 70 naming rules because of complexity of uri patterns we have there. Regex was best thing that cleaning urls framework had before so it’s really disappointing that it will no go back there… This regex was covered UUID that shows up in mutliple places in URIS, sometimes more that once so simple before after isn’t best idea. Sometimes uris are ending with UUID so we have to preserve proper order of rules to be sure that we will not merge similar uris that has something after UUID. In general for me it is huge issue.
I second this as a potential candidate for an RFE to restore previous functionality to include regexp. It’s hard enough without ability to copy rules to make some tweaks to multiple similar rules, and also it’s not available through API.
So regex was always a good option if required, and this will surely present some blockers on converting existing rules to the new format and for converting AppMon business transactions to naming rules in Dynatrace.
I’m experiencing the same problem trying to clean UUID, I try in several ways but without good results. I tried as well creating a placeholder to clean the UUID but I think there’s a bug in the “Processing steps” sections, in fact, if I place a replace rule with a leading and trailing delimiter, this is applied to all the text if the trailing delimiter is not present in the text to clean resulting in a quite impossible control of replacement. In my opinion the replace step should work as:
- Only “Leading delimiter” → Replace everything after delimiter (if delimiter exist in text)
- Only “Trailing delimiter” → Replace everything before delimiter (if delimiter exist in text)
- Leading & Trailing delimiter → Replace everything inside delimiters (only if both delimiters exist in text)
Anytime time line for the planned improvements? We too need this functionality in order to mask sensitive information like click on bank account number. Our develops cannot make the change at the code level.
I just tried to reproduce your mentioned scenario by doing the following steps in my demo environment, where I obtained the following results:
I created a “replace” placeholder with a rule like “journeyId=” in the leading part (first match) and an empty trailing part (doesn’t matter whether you pick first or last).
Hey Yair, while the new action naming is supposed to be simpler for our customers, we are currently looking into also supporting regular expressions, to make the definition of rules as dynamic as before. We will keep you updated on this!