Graylog2 / Graylog2/graylog2-server
Strange work of alerts in Graylog 3.1
- Dominant language
- Java
- Stars
- 8.1k
- Forks
- 1.1k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 217
Description
Hi
I created alert with paramaters:
- Condition Type: Filters & aggregation
- Search query: `(message:"error" OR message:"panic")`
- Search within the last: 3m
- Execute search every: 30s
- Create Events for Definition if...: Filter has results
- Notification: HTTP Notification
- Grace Period: 30s
- Message Backlog: 3
And have some problems.
## Expected Behavior
it was expected that if there are errors in the logs, every 30 seconds I will receive a notification with 3 messages
## Current Behavior
In fact, for every 1 message from source stream i have 5 events (in event stream) and also 2 hours after the problem, I received a notification about messages in the logs that were for a long time ago
source stream:

All events stream:

## Possible Solution
In my opinion, the number of events cannot exceed the initial number of messages in the logs, but in general should be less than it due to the use of the grace period and message backlog options
## Your Environment
* Graylog Version: 3.1.2
* Elasticsearch Version: 6.8.2
* MongoDB Version: 4.2
* Operating System: Ubuntu 18.04
Contributor guide
Assessment
This issue has not been assessed yet.