B
Bruce
Hello
I am designing an app to do log search efficiently. I have gigabytes
of server logs that contain all kind of information - typically I
query about a user name in a certain time span to find out what the
user actually did during that time, what errors he got etc.
I previously just used findstr across these files to do it but I am
finding it slow and inaccurate.
So, I am planning to write a platform that parses all the logs
realtime, stores the words in a database.
For example a line in the log reading "User connected to server"
would result in 4 rows in the database for each of the words, with
information about the file, time, relative location in the log among
other things.
This way, if I query for "bruce connected', I would be able to convert
it into a database query and fetch the results fairly quickly.
I have a couple of questions:
1. I am not using any standard search engine since I don't think they
index and provide the level of detail I would need. So, does my design
of using a database in the above manner sound good?
2. On top of this platform, I plan to build layers that do intelligent
search - say using business logic, it queries and finds out all users
who got errors and displays them in a UI.
I am curious to know whether there is a better approach to this.
Thanks
Bruce
I am designing an app to do log search efficiently. I have gigabytes
of server logs that contain all kind of information - typically I
query about a user name in a certain time span to find out what the
user actually did during that time, what errors he got etc.
I previously just used findstr across these files to do it but I am
finding it slow and inaccurate.
So, I am planning to write a platform that parses all the logs
realtime, stores the words in a database.
For example a line in the log reading "User connected to server"
would result in 4 rows in the database for each of the words, with
information about the file, time, relative location in the log among
other things.
This way, if I query for "bruce connected', I would be able to convert
it into a database query and fetch the results fairly quickly.
I have a couple of questions:
1. I am not using any standard search engine since I don't think they
index and provide the level of detail I would need. So, does my design
of using a database in the above manner sound good?
2. On top of this platform, I plan to build layers that do intelligent
search - say using business logic, it queries and finds out all users
who got errors and displays them in a UI.
I am curious to know whether there is a better approach to this.
Thanks
Bruce