You have made my life so easier. not a single video like this on TH-cam who would have cleared distributed tracing to me in such concise manner. Kudos to you brother. Stay blessed.
Hi Marcel Thanks a lot for sharing the knowledge one more time. I ALWAYS learn a lot from your videos And about having another video deep diving into jaeger would be great! Thanks again!
This video was super helpful! The Open Telemetry docs have been frustratingly confusing to understand and try to auto instrument the whole world without telling me what it's actually doing.
This is the issue we are facing over on my team at work as well. We are trying to adopt opentelemetry to build an end-to-end observability and resource utilization metering solution with OTEL as the backbone for the whole solution. However, the docs really aren’t that great. I found it took us many POCs of small instrumentation on little APIs to learn the flow of things and how things work. But even then, the docs are much more centered on readers who know all this observability and monitoring jargon and niche. It’s not too helpful to new engineers. I am considering getting involved with the project and specifically trying to overhaul the docs.. that would be a cool challenge.
I've been given a mandate to migrate away from Bugsnag. Bugsnag is sooo easy to implement. I only want to record errors and attach context for the error. But when I was told to use OTEL, it's like a complication explosion happened. There are like 50 different pieces, all named, all have been abstracted with their own various implementations, some auto-instrument, but that's not enough for me since we want to record OUR errors. And while tracing is cool, that's a nice-to-have for another day. Add on top of that that we use Lambda, then things get a lot more complicated. If I add the recommended lambda layer, then we exceed the 5-layer-max for a lambda function in AWS for some functions. All of our lambda functions are wrapped in my own Bugsnag wrapper and I just want to swap out the existing code with some OTEL thing that will essentially only capture errors with context. I don't want to add a context argument to all of my NodeJS functions. I only want to change the existing utils.js library in our utils layer. What do you recommend to get started?
How can get user journey from frontend to backend? I can get traces from react and nodejs but could not link it. Is jaeger good to view journeys? Plz suggest any other tool
Hello Marcel, Thanks for the nice explanation. One quick question, Can we implement context propagation using opentelemetry collector/agent of third party (e.g. Forgerock IDAM suite) product logs to Splunk or NewRelic dashboard for monitoring and analysis. Here, the third party application (ForgeRock) is not using the opentelemetry API/SDKs to handle the context propagation headers that OpenTelemetry recognizes. Please provide some example/sample reference. Thanks a lot.
Sometimes I feel like distribution of services like this adds complexity to the system for marginal benefit (different teams working on different apis) For example, why not make your playlists api and videos api monolithic and part of the same process/application?
This is the big microservice vs monolith debate. I've worked on both monolith and microservices and to be honest, it really depends on your architecture,team culture \ mindset on how you design these services. Microservices dont have to be micro or nano. a Microservice in my mind is not about the size of the service but having enough logic in there where it makes sense and defining clear boundaries. A big win in my experience is having a simplistic design with most traffic being top down (ingress to service) and not too much horizontal flows (service to service) communication. In my experience, at scale if done right, microservices can make teams insanely agile and pump out 30 - 100 deployments per day. Features can be deployed quickly, toggled on and off. You can have more people work on the platform not having to understand a large codebases, which boosts their confidence. For example, if Bob masters the videos API and understands its contract, he would be super confident to code on it without breaking other dependencies. There are a lot of wins with microservices, but in my experience it has to be done right, otherwise it can be unnecessarily complex, quite easily. And you're right, it can add a lot of complexity and there needs to be close architectural guidance to keep things as simple as possible.
@@MarcelDempers Hey, Thanks for that input.. I absolutely agree that you would need very good architectural process so things dont go haywire LOL. Yeah, I'm learning as I progress my career that a lot of times, the simplest solution wins simply because it's easy to maintain and won't cause headaches in the future LOL. Anyways, have a great day !!
@@MarcelDempers How would contracts and api be maintained among microservices? And how to track dependancies between microservices? There should be some open standards for maintaining services catalouges, contracts and dependancy graphes. Have you came across such thing in your career ?
Anyone else in this comment section new to observability and/or OpenTelemetry? I think I’m starting to center my focus in my career around this area of software but still not sure.. I’d love to create community of newbies and experts alike.
I rarely ever comment on YT but I just had to let you know how thorough, easy to follow and informative your video is.
+1 His videos are super super clear and helpful
Dude, where have you been all my life. This is the best explanation of Observability I have even seen! TY TY TY soooo much!
You have made my life so easier. not a single video like this on TH-cam who would have cleared distributed tracing to me in such concise manner. Kudos to you brother. Stay blessed.
Sir, I really glad to hear your clearly speech. Its really important to me as not native English speaker.
Let me say this: Less nonsense and 100% sense.
Crisp, clear and relatable.
Very underrated channel. This deserves lot many subscribers and followers. Thanks a lot.
This was the best video to get an understanding of tracing and instrumentation. Thank you very much!
Excellent video. Congrats, Marcel.
And please, feed more with Jaeger videos. THankss
Great content, content density. Not going to get this from reading any blog. Thanks
Hi Marcel
Thanks a lot for sharing the knowledge one more time. I ALWAYS learn a lot from your videos
And about having another video deep diving into jaeger would be great! Thanks again!
Jaeger would be awesome!
This video was super helpful! The Open Telemetry docs have been frustratingly confusing to understand and try to auto instrument the whole world without telling me what it's actually doing.
This is the issue we are facing over on my team at work as well.
We are trying to adopt opentelemetry to build an end-to-end observability and resource utilization metering solution with OTEL as the backbone for the whole solution. However, the docs really aren’t that great.
I found it took us many POCs of small instrumentation on little APIs to learn the flow of things and how things work. But even then, the docs are much more centered on readers who know all this observability and monitoring jargon and niche. It’s not too helpful to new engineers.
I am considering getting involved with the project and specifically trying to overhaul the docs.. that would be a cool challenge.
So crisp and thorugh!! Love it!
Hi, thanks for the effort and sharing. Looking forward to more in-depth videos about openTelemetry.
fantastic lesson. Great editing, illustrations, sound mix and the way things are explained.
subscribed!
Great introduction into how OpenTracing works and how to instrument your code.
great explanation, thanks! It's really helps me to undestand what tracing is.
Thank you for your explanation. You’re God sent.
Excellent production, great content from a technical perspective. Thanks!
Thanks! Very helpful one. Better than 100 pages doc!
I've been given a mandate to migrate away from Bugsnag. Bugsnag is sooo easy to implement. I only want to record errors and attach context for the error. But when I was told to use OTEL, it's like a complication explosion happened. There are like 50 different pieces, all named, all have been abstracted with their own various implementations, some auto-instrument, but that's not enough for me since we want to record OUR errors. And while tracing is cool, that's a nice-to-have for another day.
Add on top of that that we use Lambda, then things get a lot more complicated. If I add the recommended lambda layer, then we exceed the 5-layer-max for a lambda function in AWS for some functions. All of our lambda functions are wrapped in my own Bugsnag wrapper and I just want to swap out the existing code with some OTEL thing that will essentially only capture errors with context.
I don't want to add a context argument to all of my NodeJS functions. I only want to change the existing utils.js library in our utils layer. What do you recommend to get started?
This was amazing! Thank you for the time and effort you put into this.
Loves your video and hopefully topic jaeger will be on the fly soon
Thank you very much! (From Brazil)
Thanks for a great explanation of tracing and how it's implemented, much appreciated!! 🤝
Great video. A lots of userful information, very well organized and prepared and all this in less than 30 mins. Thanks a lot!
Thank you for creating such valuable resources.
So nicely explained and awesome example too , kudos!
This is amazing tutorial, kudo
Can we get the javascript files and method name details in tracing in a non node.js app?
Very informative! Very good! Could you also do Jaeger deep dive? Would really appreciate it.Thanks!
Very clean and to the point. Awesome 👏👏👏 explanation
Nice! I've been waiting for this one ;)
This is really cool. If you implement it in python specifically with grpc, it would be really really helpful and appreciated.
GOD, i repeat, you are god. What quality!
Woah! thanks for the detailed explanation
How can get user journey from frontend to backend?
I can get traces from react and nodejs but could not link it.
Is jaeger good to view journeys? Plz suggest any other tool
Cool video tutorials, I deeply understood the concept for tracing.. if u have this projects in Java, please do share. Thanks dude 👍
I've probably never wrote a youtube comment but here we are, you are awesome!
🙏🏼
Really informative thanks! But is there no other way than to manually add soans everywhere? I'd love that to happen automatically...
Great video! Can you also do a Datadog APM walkthrough?
Is there a way to auto instrument angular application using opentelemetry? Please help!!
nice explanation, thank you
Excellent ! Thank you .
great overview. your youtube video markers are a little bit off for some reason.
Hello Marcel, Thanks for the nice explanation. One quick question, Can we implement context propagation using opentelemetry collector/agent of third party (e.g. Forgerock IDAM suite) product logs to Splunk or NewRelic dashboard for monitoring and analysis. Here, the third party application (ForgeRock) is not using the opentelemetry API/SDKs to handle the context propagation headers that OpenTelemetry recognizes. Please provide some example/sample reference. Thanks a lot.
Great Content and Succinct information. Subscribed.
Thank you for this!
Good sir, could you please do a Jaeger introduction video when you get some time ? This video is great for beginners!
have you got an example, similar to this, in Python?
awsome. how do you generate the ASCII Diagram on readme files?
💪🏽 asciiflow.com
Sometimes I feel like distribution of services like this adds complexity to the system for marginal benefit (different teams working on different apis) For example, why not make your playlists api and videos api monolithic and part of the same process/application?
This is the big microservice vs monolith debate.
I've worked on both monolith and microservices and to be honest, it really depends on your architecture,team culture \ mindset on how you design these services.
Microservices dont have to be micro or nano. a Microservice in my mind is not about the size of the service but having enough logic in there where it makes sense and defining clear boundaries.
A big win in my experience is having a simplistic design with most traffic being top down (ingress to service) and not too much horizontal flows (service to service) communication.
In my experience, at scale if done right, microservices can make teams insanely agile and pump out 30 - 100 deployments per day. Features can be deployed quickly, toggled on and off. You can have more people work on the platform not having to understand a large codebases, which boosts their confidence. For example, if Bob masters the videos API and understands its contract, he would be super confident to code on it without breaking other dependencies.
There are a lot of wins with microservices, but in my experience it has to be done right, otherwise it can be unnecessarily complex, quite easily. And you're right, it can add a lot of complexity and there needs to be close architectural guidance to keep things as simple as possible.
@@MarcelDempers Hey, Thanks for that input.. I absolutely agree that you would need very good architectural process so things dont go haywire LOL. Yeah, I'm learning as I progress my career that a lot of times, the simplest solution wins simply because it's easy to maintain and won't cause headaches in the future LOL. Anyways, have a great day !!
@@MarcelDempers How would contracts and api be maintained among microservices? And how to track dependancies between microservices? There should be some open standards for maintaining services catalouges, contracts and dependancy graphes. Have you came across such thing in your career ?
Good and informative
Hi Marcel, Thanks a lot for sharing the videos, great information and the explanation too. How can we do the same tracing using Kubernetes ?
What is your VSCode theme ?
In the video its Amethyst Dark, but latest is Reloaded Dark
Awesome video!
Do you think that tracing will overtake logging?
Thanks for the kind words 💪🏽
Definitely not, since they have two entirely different goals.
Great video, thanks!
Thanks Marcel!
Thanks Marcel
Very useful
Thank you
nice and informative.
thanks a lot =)
Catch an exclusive interview with Binance’s CEO to get the scoop on future developments
Like first, watch after ;)
YesJaeger video please!
Anyone manage to achieve this in .net?
Peace!
Anyone else in this comment section new to observability and/or OpenTelemetry?
I think I’m starting to center my focus in my career around this area of software but still not sure.. I’d love to create community of newbies and experts alike.
great!
It it requires adding a ctx parameter to each function that needs to be traced , its not going to happen.
those arms!!!
lekker SA accent ;)
Like 666!