[Anonymize Question] Is there a way to anonymize url in Traceback?

[xiao623]

I need to set https://<user>:<password>@xxxxx.com:9200 for my secured elasticsearch cluster and i also konw that we can use anonymize_url when calling logger.xxx().

However, I had found that the message <user>:<password> still could be known from the Traceback.

For Example,
https://github.com/chaoss/grimoirelab-elk/blob/5db69f80e0d436ccf9be2403ae1bf98529337545/grimoire_elk/elastic.py#L181-L182

 res = self.requests.put(url + '?refresh=true', data=bulk_json, headers=headers)
 res.raise_for_status()

Here, url would be https://<user>:<password>@xxxxx.com:9200.
But if there a wrong status_code in raise_for_status, the messge would be :
https://github.com/psf/requests/blob/1e5fad7433772b648fcbc921e2a79de5c4c6be8b/requests/models.py#L954

http_error_msg = u'%s Client Error: %s for url: %s' % (self.status_code, reason, self.url)

and raise HTTPError with self.url which contains the message <user>:<password>.

Like #309 (comment)
requests.exceptions.HTTPError: 400 Client Error: Bad Request for url: https://XXX:XXX@e69eb515c21ebbcaf026c5d04dc49f1a.us-east-1.aws.found.io:9243/.kibana/doc/dashboard:Overview.

I had tried to use anonymize_url in the method feed_backend in elk.py but the message <user>:<password> still be shown in Traceback

https://github.com/chaoss/grimoirelab-elk/blob/5db69f80e0d436ccf9be2403ae1bf98529337545/grimoire_elk/elk.py#L178

-           error_msg = "Error feeding raw from {} ({}): {}".format(backend_name, backend.origin, ex)
+          error_msg = "Error feeding raw from {} ({}): {}".format(backend_name, backend.origin, anonymize_url(str(ex)))

[xiao623]

• Is that exec_info=True necessary? Maybe we could remove it.

    logger.error("Error enriching raw {}".format(ex), exc_info=True)

• For other logger.xxx(...., ex), the ex might expose sensitive info because of raise_for_status.
  And when we config the logger there is a useful solution : https://relaxdiego.com/2014/07/logging-in-python.html