snitch 0.1

JSON log formatter and Sentry client

Utilities for working with [Sentry](https://github.com/getsentry/sentry). We got into trouble using the in-app Sentry client [raven-python](https://github.com/getsentry/raven-python) client with high volumes so we decided to take the HTTP posting to Sentry out of the app itself. Snitch helps here.

log2json

is a stdlib log formatter that will output log record in Sentry’s JSON format. You can use any handler, but we use it with a file handler so the app can always quickly dump the log record and does not have to wait for an HTTP post.

snitch

monitors a logfile for new lines (like tail but skips all existing content) and sends each new line in a HTTP POST request to the Sentry API. Snitch reloads the log file every half minute or so. On one hand, this is a workaround to deal with log rotation, on the other hand, this deals with high log volumes by just dropping everything and resynching with the present. I mean, when it reopens the file it effectively skips everything that was already logged to the file but not yet processed because it continues at the end of the file after a reopen. This is what we want. Also, snitch does a time.sleep(.1) between all POSTS as a rate-limiter.

alternatives

I haven’t looked into using UDP instead of TCP for sending the messages. This may solve problems on the sending host, but still has the potential of flooding Sentry.