News for hackers. Stuff that matters.
Re:Show HN: Errorship, use datadog as an error tracker by tyingq on Sunday April 19, 2020 @07:53AM (#22915237)
Can you add a little more color commentary? I haven't used datadog, but they seem clear that there's both a logging and metrics (apm) side to their product. Is the logging side neglected somehow?

I see, for example, docs on how to log from Python: https://www.datadoghq.com/blog/python-logging-best-practices...

Re:Show HN: Errorship, use datadog as an error tracker by RabbitmqGuy on Sunday April 19, 2020 @08:02AM (#22915287)
Hi.

> but they seem clear that there's both a logging and metrics (apm) side to their product.

yes they have those two sides. what they do not have is an error tracking side; something like sentry[1]. errorship exists to bring the error tracking side into datadog.

The logging side of datadog is like splunk or elasticsearch, and their APM side is like newrelic. errorship is bringing the error tracking side.

some of the analogies may not apply 100%, but I hope I'm making sense.

1. https://sentry.io/

Re:Show HN: Errorship, use datadog as an error tracker by tyingq on Sunday April 19, 2020 @08:07AM (#22915318)
That's helpful. Basically zero-config central logging of exceptions/stderr/whatever?
Re:Show HN: Errorship, use datadog as an error tracker by RabbitmqGuy on Sunday April 19, 2020 @08:29AM (#22915448)
Yeah.

But also, errorship will still capture(and send to datadog) any exceptions that you may have forgot to capture or log.

Re:Show HN: Errorship, use datadog as an error tracker by EdwardDiego on Sunday April 19, 2020 @08:20AM (#22915389)
I'm guessing that using the DD event stream rather than purchasing logging support could be a cost saving, although not sure how DD would feel about that.
Re:Show HN: Errorship, use datadog as an error tracker by RabbitmqGuy on Sunday April 19, 2020 @08:37AM (#22915501)
errorship is not using the datadog eventstream to send all logs to it. Rather, errorship just sends exceptions, their stack traces(and a few other metadata like hostname) to the eventstream.

> although not sure how DD would feel about that

The datadog documentation[1] says; "An event represents any record of activity noteworthy for engineers"[1]. We certainly believe that exceptions that occur in engineer's applications are noteworthy activities and as such belong to the eventstream.

1. https://docs.datadoghq.com/events/

Re:Show HN: Errorship, use datadog as an error tracker by EdwardDiego on Sunday April 19, 2020 @08:43AM (#22915527)
> errorship is not using the datadog eventstream to send all logs to it.

Yep, noted that, and I like how you preserve syntax highlighting in the stack traces :)

Out of curiosity, I've only ever dabbled in Datadog, is there any limit on event stream size or retention? I couldn't find anything mentioned in the docs.

Re:Show HN: Errorship, use datadog as an error tracker by RabbitmqGuy on Sunday April 19, 2020 @08:58AM (#22915595)
I had also noticed the lack of mention in the docs, so I reached out to datadog sometime last year. Their response was that it doesn't have any limits or quota
Re:Show HN: Errorship, use datadog as an error tracker by tyingq on Sunday April 19, 2020 @09:45AM (#22915840)
Interesting. Is there some disincentive that discourages it? Is it hard (or costly) to query, for example...compared to paying specifically for logging?

Seems like a big loophole.