Is there a system for analyzing arbitrary logs from a database, BugTracker?

Question

Monitoring

Good afternoon, colleagues.
The essence of the task is as follows:
1. There is a single document storage system, an electronic archive
2. Several accounting systems send their documents to the storage system
3. IS receives a response from the service, in case of failure it receives an error text
4. IS fixes an error in the local log I
plan to add a function additional recording of the error text in a separate database. A service is needed to analyze this database with the ability to inform in case of triggering certain triggers.
In fact, you need to visualize the database, group by events, select by type of IP, group by employee, and so on and so forth.

Answer 1

To store partially or unstructured logs (text), many people usually use the ElasticSearch DBMS in conjunction with the name ELK, in which LogStash is like a crusher of different types of logs. In this case, there is a selection of various data structures that eventually end up in the DBMS. This allows you to search by various filters through the web interface.
Or just take the Loggly service for storing logs and a search interface.
If you need an alert mechanism with triggers, then you should store numerical metrics in a Prometheus-type time series DBMS and configure AlertManager for groups of different subscribers.
To do this, each service must periodically return HTTP metrics along the /metrics path upon request from Prometheus. The language for statistical selections is quite flexible and covered all our needs.
Let's say you need to notify if the load on the service has exceeded 30% compared to the previous week / month.
Prometheus is not good for storing logs, but good for metrics!
Perhaps the Amazon CloudWatch service allows all of the above. Didn't check.

Answer 2

Thanks, I'll take a look at the suggested solutions.