Fetch logs from RDS postgres instance and use them with pgbadger to generate a report.
Fetches RDS log files and analyzes them with pgBadger.
Make sure your credentials are set in the ~/.aws/credentials file. Also, you can set a region in the ~/.aws/config file, so passing region option to the script is not needed. Last but not least, make sure you have pgbadger installed and reacheable from your $PATH.
You will have to configure your database parameter group.
First of all, ensure log_min_duration_statement is set to 0 or higher, else you won’t have anything to be parsed.
Then you must enable some other parameters to get more information in the logs.
Also make sure lc_messages is either at engine default or set to C.
For further details, please refer to Dalibo’s pgBadger documentation.
You can install it using pip:
$ pip install rdspgbadger
To build a pgbadger report, just run the following (replacing instanceid by your instance ID):
$ rds-pgbadger instanceid
Only the Instance ID is mandatory, but there are also other options you can use:
- -d, –date : by default the script downloads all the available logs. By specifying a date in the format YYYY-MM-DD, you can then download only that day’s logs.
- -r, –region : by default the script use the region specified in your AWS config file. If none, or if you wish to change it, you can use this option to do so.
- -o, –output : by default the script outputs log files and reports to the out folder. This option allows you to change it.
- -n, –no-process : download log file(s), but do not process them with pgBadger.
- -X, –pgbadger-args : command-line arguments to pass to pgBadger
- –assume-role : By specifying a role you can use STS to assume a role, which is useful for cross account access with out having to setup the .config file. Format arn:aws:iam::<account_id>:<role_name>
In spite of the great work of askainet, AWS API seems to be too instable, and sometimes download of big log files can fail. In such case retrying a few minutes later seems to work.
see pull request 10
For any request, feel free to make a pull request or fill an issue on Github.