After a delay of over a 12 months, an open supply code contribution to allow the export of knowledge from Datadog’s Utility Efficiency Monitoring (APM) platform lastly bought merged on Tuesday into a group of OpenTelemetry parts.
The rationale for the delay, in line with John Dorman, the software program developer who wrote the Datadog APM Receiver code, is that, a few 12 months in the past, Datadog requested him to not submit the software program.
On February 8 final 12 months Dorman, who goes by the identify “boostchicken” on GitHub, introduced that he was closing his pull request – the git time period for programming code contributed to a mission.
“After some consideration I’ve determined to shut this PR [pull request],” he wrote. “[T]listed below are higher methods to OTEL [OpenTelemetry] assist w/ Datadog.”
Members of the open supply group who’re centered on utility monitoring – amassing and analyzing logs, traces of app exercise, and different metrics that may be helpful to maintain functions operating – had questions, claiming that DataDog prefers to lock clients into their product.
Shortly after the submit, Charity Majors, CEO of Honeycomb.io, a rival utility monitoring agency, wrote a Twitter thread elaborating on the advantages of OpenTelemetry and calling out Datadog for under supporting OTEL as a one-way avenue.
“Datadog has been telling customers they’ll use OTEL to get knowledge in, however not get knowledge out,” Majors wrote. “The Datadog OTEL collector PR was silently killed. The one that wrote it seems to have been pressured into closing it, and nothing has been proposed to exchange it.”
Habits of this type can be inconsistent with the objectives of the Cloud Native Computing Basis’s (CNCF) OpenTelemetry project, which seeks “to supply a set of standardized vendor-agnostic SDKs, APIs, and instruments for ingesting, remodeling, and sending knowledge to an Observability back-end (i.e. open supply or business vendor).”
That’s to say, the OpenTelemetry mission goals to advertise knowledge portability, as a substitute of hindering it, as is widespread amongst proprietary software program distributors.
The smoking hound
On January 26 Dorman confirmed suspicions that he had been approached by Datadog and requested to not proceed together with his efforts.
“I owe the group an apology on this one,” Dorman wrote in his pull request thread. “I lacked the braveness of my convictions and when push got here to shove and I needed to make the laborious selection, I took the simple means out.”
“Datadog ‘requested’ me to kill this pull request. There have been different members from my group current that allow me know this reply will probably be a ‘okay’. I’m positive I may have stated no, in the intervening time I simply could not fathom opening Pandora’s Field. There you may have it, no NDA, no stack of money. I left the code hoping somebody may stick with it. I used to be keen to provide [Datadog] this code, no strings connected so long as it moved OTel ahead. They declined.”
He added, “Nonetheless, I instructed them in the event you do not assist OpenTelemetry in a significant means, I’ll begin sending pull requests once more. So right here we’re. I really feel I’ve given them sufficient time to do the fitting factor.”
Certainly, Dorman subsequently re-opened his pull request, which on Tuesday was merged into the repository for Open Telemetry Collector parts. His Datadog ARM Receiver can ingest traces within the Datadog Hint Agent Format.
Coincidentally, Datadog on Tuesday revealed a blog post titled, “Datadog’s dedication to OpenTelemetry and the open supply group.” It makes no point out of the alleged request to “kill [the] pull request.” As an alternative, it enumerates varied methods during which the corporate has supported OpenTelemetry not too long ago.
The Register requested Datadog for remark. We have not heard again.
Dorman, who presently works for Meta, didn’t reply to a request for remark. Nonetheless, final week, via Twitter, he credited Grafana, an open supply Datadog competitor, for having “formally sponsored” the work and for mentioning that Datadog “refuses to assist OTEL in significant methods.”
The OpenTelemetry Governance Committee for the CNCF offered The Register with the next assertion:
“We’re nonetheless making an attempt to make sense of what occurred right here; we’ll touch upon it as soon as we have now a full understanding. Regardless, we’re comfortable to assessment and settle for any contributions which push the mission ahead, and this [pull request] was merged yesterday,” it stated. ®
