I need to log REALLY BIG xml payloads (soap request/response messages). In general, most of them are not that big and log to Seq just fine. However, some of them are as large as 1.5MB and they are not logging. Looking at the ingestion log I see the message: Rejecting payload from "::1" due to oversized event; first 1024 chars of 1506249 Is there a way to increase the maximum event size?
Posted by Patrick Steele 9 months ago
Seq is exposed as localhost:5341, but I want it to be exposed as seq.domain.com. I'm running on a VM in azure. I tried adding it to the URI to the seq.json, but still couldn't access it (added to hosts file). I also tried doing a reverse proxy through IIS, also failed. Really want to get it to work :)
Posted by Dan Shapir 9 months ago
So I've setup a separate VM in azure and bound port 443 to seq, from my local machine I can write to it just fine since I've added my own CA to my machines trusted root cert store. Once I deploy to azure I can't log. I'm guessing it's since I'm using a self-signed cert and I can't install certs to the trusted root store for azure websites. And I really want my logging application to run in an azure website. To the question: is it possible to configure serilog to not validate the certificate chain? Perhaps point out my own request handler where I set the validation callback to always return true?
Posted by Ajden 10 months ago
I've tried to implement the Dynamic Level Control used your example: Log.Logger = new LoggerConfiguration() .MinimumLevel.ControlledBy(levelSwitch) .WriteTo.Seq("http://localhost:5341", apiKey: "yeEZyL3SMcxEKUijBjN", controlLevelSwitch: levelSwitch) .CreateLogger(); But controlLevelSwitch is not an argument of Seq Sink. All libraries are updated. Any suggestions?
Posted by Oleg Gryaznov 10 months ago
I'm using an App which accepts a signal. Only one signal can be specified in the app, but I'd like to use a combination of signals. I could create a new signal with a set of filters which duplicate the filters in the other signals, but it would be nice to not have to redefine filters which exist elsewhere. Or would it be better to modify the App so it accepts multiple signals?
Posted by Chris 10 months ago
I know, that currently the main issue of Seq is that it can't work in a cluster. And also if we have a lot of logs in a minute we could have troubles. I think one of the issues is that you are using flat files as a database. So we can't use Seq as a cluster on different servers. And also searching through files is much longer, than through database. Are you planning to have some changes with this in future versions? Thanks.
Posted by Oleg Gryaznov 10 months ago
Hello, I have a system setup with 2 separate applications communicating to each other through a message queue on the same machine. Requests come into App A -> MQ -> App B -> MQ -> App A. This happens in less than a millisecond. When I look at the event stream in Seq (3.1.16) the messages are displayed (from the bottom up): App B App A App A This should be the other way around, any idea how I can check the precision of the time stamps for log messages being received by Seq? Given the apps are on the same machine the timestamps should be accurate.
Posted by Laurence Evans 10 months ago
When looking at errors ... mostly you get ALOT ... of the same when doing a search on some property ... when afterward looking at the data ... you might want to filter out even more ... There it would be nice to have a "Add To Query", instead of the "Find", which just replaces the whole query ...
Posted by Mikael Syska 10 months ago
Is there any chance to get something analogous to RegEx back references available in Seq syntax? eg. ($45612) && CorrelationId == 1.CorrelationId <- in this case 1 is the implied back reference. The problem it helps solve: This is particularly useful for troubleshooting warnings, errors and workflows.
Posted by At 11 months ago