I would like to create a dashboard chart that displays the number of each type of event at an hour interval. My query looks like this: select count(*) from stream where Application = '<application>' group by @EventType, time(1h) When I create the graph based on this query, it changes the interval to 5s: select count(*) as col0 from stream where Application = '<application>' group by @EventType, time(5s) limit 10000
Posted by Dorian Dechant about a year ago
On your mission page you state that "(Seq is) not an "ops" tool that lives only in production environments; it's light enough to live alongside your running .NET apps from the very first day of development.". This is exactly what I have been using seq for, as an alternative to rolling log files, however half of my .Net stack is dotnetcore and runs on linux. So, is there any chance that we will be able to run seq on linux boxes in the future?
Posted by Bill Haggerty(So user Theyouthis) about a year ago
I noticed that our license will cease to get updates soon. Does this mean that we can install all updates which were released before that date, or will we even be prevented from updating to versions which were released before the expiration of support?
Posted by Ryan Adler about a year ago
We have installed Seq on of the VM and started windows service. We can see the seq dashboard. As soon as we start our other process which uses serilog to send messages at 5341 port, Seq service just stops without much error in Event manager. If use visual studio to debug the process of seq we don't get any error. Please help!
Posted by Salil Shah about a year ago
For some reason, there seems to be two users with the same name in Seq which makes it impossible to login. I've tried using the `seq unlock --username=myusername --password=mypassword` but I get the error `Could not reset the password: Sequence contains more than one matching element`. This is kind of critical since this is our production Seq machine. How can I reset the users in Seq and login as administrator again?
Posted by Patrik Svensson about a year ago
Hello, I've got a setup where I've got 2 named instances of Seq on a single server and I want each of them to have a seperate SSL certificate. But I don't know how to do this, it looks like when I set the SSL certificate it is used for both instances. My setup is as follows: 1. Named instance with url like: seq01.company.com 2. Named instance with url like: seq02.company.com I would like to set a different SSL certificate for each instance, otherwise I'll get a certificate invalid error on one of the instances. How can I achieve this? The bind-ssl command doesn't seem to allow for a named instance parameter. Kind Regards, Matthijs
Posted by Matthijs Woolderink about a year ago
While evaulating Seq on a test server, we've twice seen the Seq service have not come back automatically after a server reboot. The only strange thing in the Event logs is both times the system time was changed to a couple hours back in time. We've tried to repro this locally, but haven't found a way to actually change the system time to confirm ("Some settings are managed by your organization"). Thus reporting this in hope somebody else can repro or confirm. Per stability troubleshooting steps: Seq v3.4.20 - The Services console lists "Seq Event Server and Web UI" and should start automatically. OS: Windows Server 2016 Standard .NET version: 4.6.01586 (60632) Hardware: VMWare, Intel(R) Xeon(R) CPU, E5530 @ 2.40GHz, 2394 Mhz, 2 Core(s), 2 Logical Processor(s), 8GB RAM + 9.25GB virtual Settings > Diagnostics: - API Endpoint events: 0/0 0/0 0/0 - Storage: The event store is arranged in segments from 2017-05-17 to 2017-06-20. - Out of 35 recorded days, 35 days totalling 62711 events are cached in RAM. - Uptime 0 days - Working set (bytes) 231411712 - Threads 25 - User thread pool threads available 32765 - IOCP thread pool threads available 999 - Memory utilization 25% - Free space (bytes) 70923096064 - Sample system memory 28 minutes ago - Manage segment cache resources 28 minutes ago Tail of log file from when the server did not come back up: 2017-05-19 14:37:04.933 +01:00 [Information] Received OS shutdown notification 2017-05-19 14:37:04.933 +01:00 [Information] Stopping HTTP listener 2017-05-19 14:37:04.933 +01:00 [Information] Signalling for 2 background worker(s) to complete 2017-05-19 14:37:06.261 +01:00 [Information] Stopping 10 scheduled tasks 2017-05-19 14:37:06.261 +01:00 [Information] Stopping broadcast channel 2017-05-19 14:37:07.173 +01:00 [Information] Seq stopped cleanly Event Viewer > System Log lists the following two events: Date: 19/05/2017 13:21:47 The Seq service failed to start due to the following error: The service did not respond to the start or control request in a timely fashion. Date: 19/05/2017 13:21:47 A timeout was reached (30000 milliseconds) while waiting for the Seq service to connect. The only odd thing I can see is the server clock being adjusted around the same times, which is known to potentially cause issues. It caught my eye because the timestamps in the logs/event viewer don't match up (besides UTC/BST confusion), and there is this system log event timestamped 14:40:47: The system time has changed to 2017-05-19T13:40:47.094288300Z from 2017-05-19T12:23:40.322019000Z. Change Reason: An application or system component changed the time. Last time Seq did not come back, on 11. may, there was a similar "changed the time" event around when the server restarted (around 02:00-04:00 in the morning). Any ideas?
Posted by Anders about a year ago
When configuring listener through the GUI there is a field for SSL thumbprint How do I specify it from command line? For example, if I want to change the port I can use this command: seq config -k api.listenUris -v https://localhost:443 But SSL does not work after seq restart. Looks like it is missing the thumbprint as SSL connection cannot be established The bind-ssl command involves "seq uninstall" (https://docs.getseq.net/docs/ssl). Is there a way to configure HTTPS and specify thumbprint without running seq uninstall? The main idea is to be able to switch the listener without uninstalling/reinstalling - just change config and restart. Is it possible?
Posted by Peter 2 years ago
Is there any way to force old events to be cached into memory? I have a 60gb database for 21 days of data. I want to get some statistics from errors from 1-2 weeks. I upgraded my AWS VM to a 32Gb instance, but even though I'm querying the previous week, I always get a timeout during the graph generation, and the Seq process won't go over 6Gb, sometimes even drops below. Is there a setting I can change to tell it to keep loading data into memory to speedup these queries and get these graphs?
Posted by Natan 2 years ago
Hello, we are logging every http request with domain account of user in a property. Currently we have a query with total request count on dashboard and it works fine. But we were unable to create a query to show number of distinct active users over period of time. Something like this would be needed: select count(*) from (select distinct(UserName) from stream where @EventType = 0x11111111 group by time (5m)) Or is there any other way to achieve this without a nested select? Regards, Vladimir
Posted by Vladimír Mihok 2 years ago