[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freedesktop-dbus
Subject:    OpenTelemetry tracing of Dbus calls
From:       Umut Tezduyar Lindskog <umut () tezduyar ! com>
Date:       2022-11-30 7:53:02
Message-ID: CAFKnvs4hagCmeDbxCU+YsobqX1mNUK6X9iaGaqBn1P7p4SK9LA () mail ! gmail ! com
[Download RAW message or body]

Hi,


We are looking into a standardized way of distributed tracing and
OpenTelemetry standard has been a winner so far. To be able to do a
distributed tracing a context needs to be passed from node to node,
application to application, process to process. This is called "Context
Propagation" in their terms. There are standard & built-in ways of
propagating calls in HTTP, grpc etc without changing the APIs.


We have network services that have a web API layer and a Dbus based service
layer. API talks to service and service might be talking to other services.
We can trace all the way from the cloud to the web API layer with minor
code additions. I believe OpenTelemetry tracing at the DBus layer would be
very valuable for the overall picture for similar use cases.


Has there been any discussions about OpenTelemetry, integration of context
propagation to Dbus. If not, what are the suggestions regarding tracing
Dbus APIs. Is there a way to hook in lttng or ePBF to dbus protocol (or to
the daemon) and convert their signals to OpenTelemetry traces?


Thanks in advance for the feedback

Umut

[Attachment #3 (text/html)]

<div dir="ltr"><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US">Hi,  </span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US"><br></span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US">We are looking into a standardized way of distributed tracing and \
OpenTelemetry standard has been a winner so far. To be able to do a distributed \
tracing a context needs to be passed from node to node, application to application, \
process to process. This is called "Context Propagation" in their terms. There are \
standard &amp; built-in ways of propagating calls in HTTP, grpc etc without changing \
the APIs.</span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US"><br></span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US">We have network services that have a web API layer and a Dbus based \
service layer. API talks to service and service might be talking to other services. \
We can trace all the way from the cloud to the web API layer with minor code \
additions. I believe OpenTelemetry tracing at the DBus layer would be very valuable \
for the overall picture for similar use cases.  </span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US"><br></span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US">Has there been any discussions about OpenTelemetry, integration of \
context propagation  to Dbus. If not, what are the suggestions regarding tracing Dbus \
APIs. Is there a way to hook in lttng or ePBF to dbus protocol (or to the daemon) and \
convert their signals to OpenTelemetry traces?  </span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US"><br></span></p><p class="MsoNormal" \
style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
style="font-size:11pt">Thanks in advance for the feedback</span></p><p \
class="MsoNormal" style="margin:0cm;font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)"><span \
lang="EN-US">Umut</span></p></div>



[prev in list] [next in list] [prev in thread] [next in thread]