dog_stats_api doesn't encode pipe char

[yannmh]

Issue by clofresh
Friday Feb 21, 2014 at 23:23 GMT
Originally opened as https://github.com/DataDog/dogapi/issues/81

---

| is used by the statsd protocol to delimit the fields, so if you use it in a tag, you'll lose whatever's after the pipe

[jirikuncar]

@DataDog/agent-core should we escape or remove | in tag?

Is there any documentation which specifies valid tags? I have only found https://docs.datadoghq.com/developers/metrics/dogstatsd_metrics_submission/?tab=python#metric-tagging.

see also #517

[arbll]

@jirikuncar https://docs.datadoghq.com/tagging/#defining-tags defines valid characters for tags. AFAIK statsd/dogstatsd do not specify how to escape characters that conflict with the protocol or are invalid. It probably make sense to either strip them or replace them by a valid character like _.

In any case let's make sure we align with other clients if they have logic to escape tags / metric names.

[jirikuncar]

@arbll thank you for the link. It doesn't say how the tags should be encoded in case of dogstatd. We can add a validation, but | seems like a valid character.

[arbll]

@jirikuncar | is invalid for all datadog tags, so it's also invalid for dogstatsd.
Again, I don't think dogstatsd/statsd escaping is defined anywhere. See statsd/statsd#585 for example

[jirikuncar]

@arbll so basically a tag has to match ^[A-Za-z][A-Za-z0-9_\-:,/]{0,199}$ after re.sub('[^A-Za-z0-9_\-:,/]', '_', tag), right?

---

Playground: https://regex101.com/r/yJ8qva/1/tests

[github-actions]

Thanks for your contribution!

This issue has been automatically marked as stale because it has not had activity in the last 30 days. Note that the issue will not be automatically closed, but this notification will remind us to investigate why there's been inactivity. Thank you for participating in the Datadog open source community.

If you would like this issue to remain open:

1. Verify that you can still reproduce the issue in the latest version of this project.

2. Comment that the issue is still reproducible and include updated details requested in the issue template.

[zippolyte]

closed in #517

The library will now sanitize tags before sending them