I chose Cilium, but Cilium was acquired by Cisio. How does this affect Cilium's growth? In my opinion, companies acquired by Cisio tend to be "passive," typically AppD.
I guess it's too early to tell. The only thing I can say, for now, is that both companies are focused on networking so one acquiring the other makes sense. We are yet to see whether logical sense will apply to Cisco.
I really like your videos, I believe it would be good if you can make a video with a list of the best Open Source tools to build a full developer platform. Using just a git repo and a bunch of bare metal servers / VMs. A poors man solution sort of say. Thanks for sharing!
Great idea with OS projects! If you find one project interesting you can always jump in and even raise a PR! How guys do search for such interesting projects like cdk8s or mentioned komodor? Conferences? Reddit?
@mrcaique1025 I'm fortunate to speak in many events and hang with quite a few folks from other companies and projects so i get a lot of info first hand. Still, that's a small part of the tools i discover. Most come from random searches and encounters. Most people try to find tools that solve specific problems they have. I tend to ignore that and explore at least one tool or project every week without any objective. Once i do spend enough time, i discard it or, in rare cases, see that i can benefit from it and adopt it.
Great video as usual! Viktor really knows how to dive into tools from a practitioner's perspective. The only thing is, I'm not 100% on board with his take on Backstage. It's no secret that Viktor has his own unique flavor of opinions about Backstage as it stands. Can't wait to see what he thinks when the new declarative approach to plugin development hits GA!
For observability a scalable Prometheus option could be an OTEL Target Allocator and Grafana's MIMIR as a backend. Additional benefit OTEL TA also respects Prometheus Operators CRs! Which is helpful for service discovery!
Hey @powersurge5576, full disclosure here, I work for groundcover. To your point, just wanted to mention that we use Victoria Metrics for the metrics backend which scales great for both short and long-term metric storage. We also use oTel collector for ingesting logs and traces, however we use a VM agent for the metrics ingestion as it is more efficient and lightweight. We have a really generous free tier - feel free to check us out. We’d love your feedback :)
I like your video and very helpful to me. , the one which I felt missing is Backup and recovery management in Kubernetes. It would be nice if you could also suggest a tool on that for 2024
I'm shocked for Database section there was not any mention of Flyway or Liquidbase. These do the necessary work of deploying SQL queries which is great. However, I do appreciate the mention of Atlas Operator. It's a cool operator that doesn't get enough mention.
Given that more and more apps are running in kubernetes it makes a lot of sense to bundle them with schemas in s way native to Kubernetes. That's why I prefer atlas operator. Flyway and Liquibase are great but did not yet, as far as I know, do the work of making a kubernetes CRD.
@@DevOpsToolkit Agreed! Would love to see more K8s Operators down the road. Bitnami's Sealed Secrets was mentioned and it has a special place for my team. Helps with our deployments of secrets in conjunction with AWS IRSA usage in AWS EKS. We use Bitnami Sealed Secrets as a backup just in case AWS STS has issues.
That's the chicken or egg type of a problem. You can use a managed service like Upbound for control plane clusters. If you prefer a self-managed solution, there are two options that come to my mind. One is to start a local cluster with Crossplane and use it to create the first "real" control plane cluster. Once it's done, you can move Crossplane and that cluster definition there and, from there on, that "real" cluster would manage itself (apart from managing everything else). Alternatively, you can create that "real" cluster any other way (e.g., eksctl, hyperscaler console, etc.). Crossplane is primarily used to manage resources at scale so one of many clusters (not counting databases and everything else) is typically not an issue. P.S. I assumed you asked about Crossplane. Please let me know if that's not the case and I'll adjust my answer.
Thank you very much for amazing content as usual :) Could you please elaborate on the infrastructure and service management using kubernetes ? Do you have a link or a video that you made about it ? I understand that its a concept and I wondered if you have some practical examples of using it :)
Interesting, for security I feel like trivy-operator is the more complete product at the moment, although it is nice that kubescape is now part of CNCF
Hey victor, i know its been 9 months since this video and the landscape can change a lot in that time 😂😅 i was wondering if these tools are appropriate for a small team (5 devs) to manage self-managed servers as we operate in regions where there are NO cloud support
The landscape did not change much since then. Also, i will be making a new one around the beginning of the next year. Unfortunately, I cannot day whether they are appropriate for such a small team. They are probably not. It might be an overkill to use kubernetes (many of those are related to kubernetes). I would go as far as to say that self-managed servers are not appropriate either. If there are only five devs in total you probably want to focus on developing whatever you're working on and not on managing infra and creating your own services. You should be using Google cloud run, azure Container Apps, fly.io, ir something similar. On the other hand, if you do have to self manage and if you do have a decently large scale (which you did not mention so i don't know whether that's the case), then kubernetes is the way to go.
There was no specific reason for not mentioning elastic other than trying to be brief. I also did not mention DataDog and Dynatrace and many others. I tried to mention only a top few, but that does not mean that there are no other good solutions (like Elastic).
You need a backup / mobility option for your managed and Kubernetes based databases…. Native operator doesn’t cut it if you have more than one database
I'd wary using kubernetes control plane for infrastructure. Heck, I even seen DNS servers on top of ESXi VMs, where the hosts relied on remote storage, which depended on DNS in the first place. A kubernetes cluster can die spectacularly in a real world environment, especially if it is provided by a cloud provider. To me infrastructure is critical enough not to rely on a platform with too many abstraction layers.
I like the idea, even though that will get me into even more trouble. How about this. A video near the end of a year with droppings, and a video around the beginning of a year with additions?
Hi, I have 8 years of experience as a programmer, I have been working as DevOps for 2 years? What technologies do you recommend learning about? With AWS I have 4 certificates, with Kubernetes I have CKAD. Overall, my stack is mainly AWS, Linux, Gitlab CI/CD, bash, Terraform and Ansible to a lesser extent, plus python, but I do very little development in it. I mainly use Zabbix for monitoring.
@@DevOpsToolkit I forgot, in the project I use Skaffold and ArgoCD. For monitoring Prometheus + Grafana , and Longhorn for storage, sometimes I will write a small HELM package, but I prefer someone else's :)
@@DevOpsToolkit The scale is not huge, so sometimes I have to go back to documentation, e.g. Ansible. As an AWS Community Builder, I have the opportunity to play around with more expensive services on AWS
Victor, could you please give an advice, how to create local development environments for Dev Teams? its a pain in the ass)) lets say, we need some Cloud services and DBs, containers accessible from the internet etc.
@@DevOpsToolkit i think, i will use local K8s (Kind) and DB instances and other services in cloud deployed by Terraform for each developer. Or maybe there are some tools like Localstack is better ? Hvala puno
Hello, I live in Korea and my job is DevOps engineer. I'm enjoying your RUclips. Can I summarize your RUclips translation into Korean and upload it to my blog? It's not for profit. It's just my hobby. I'm studying various new tools while watching your RUclips, and I want to share them with my close friends. Furthermore, I would like to share it with IT engineers in Korea. Will it be possible? And if you don't mind... Can I also post it on the blog of my company?
Thanks for another great video, we really needed a list after a lot of videos! I agree with most of the choices you made, except for maybe databases (for postgres I'd evaluate crunchy Pg) and pipelines (I'd rather use gitlab for now)
The beauty of Dagger is that it does not replace pipelines. You still need something like GitLab to run it. I, for example, use GitHub Actions but inside them there is invocation of Dagger. That alone would not be enough for me to use it but then there is the ability to run the same locally as well. In other words, instead of putting Shell scripts inside pipelines, I put Dagger code.
Thanks for the update on this, I think I should have specified I'm using remote development everywhere I can. Among the other problems it solves it makes running pipelines feel like they were "local' as they are just like one of the other services. But I watched your video about it and still think dagger is a good idea for people who want to have the same behaviour in different environments
@IvanRizzante oh yeah. My preference is to use remote environments as well and in those cases dagger does not really provide much, if any value. Unfortinately, remote environments adoption is slower than i hoped and that's one of the predictions i had in the past and failed.
The picks are mostly among those I invested heavilty in 2023. There's no Ingress that sparked my attention lately. Now it's mostly about Gate API implementations.
k8 is what happen when there is a lack of leadership on the project, and they simply accept PRs about any feature, in 2030 k8s will become an OS that you can install from an iso file and then use Microsoft Office
What are your top choices?
I chose Cilium, but Cilium was acquired by Cisio. How does this affect Cilium's growth? In my opinion, companies acquired by Cisio tend to be "passive," typically AppD.
I guess it's too early to tell. The only thing I can say, for now, is that both companies are focused on networking so one acquiring the other makes sense. We are yet to see whether logical sense will apply to Cisco.
I was wondering where you think Hubble fits in since you said cilium and ebpf are doing great things ( i also agree).
@@kungfu71186 Hubble is great. I should have included it into the list explicitly.
Pulumi
Thanks Victor for choosing the kubescape! I promise we will continue with the good work also in 2024💪stay tuned!
I really like your videos, I believe it would be good if you can make a video with a list of the best Open Source tools to build a full developer platform. Using just a git repo and a bunch of bare metal servers / VMs. A poors man solution sort of say. Thanks for sharing!
I did such a video, but it wasn't limited only to open source tools. I'll add it to my TODO list...
Great idea with OS projects! If you find one project interesting you can always jump in and even raise a PR! How guys do search for such interesting projects like cdk8s or mentioned komodor? Conferences? Reddit?
@mrcaique1025 I'm fortunate to speak in many events and hang with quite a few folks from other companies and projects so i get a lot of info first hand. Still, that's a small part of the tools i discover. Most come from random searches and encounters. Most people try to find tools that solve specific problems they have. I tend to ignore that and explore at least one tool or project every week without any objective. Once i do spend enough time, i discard it or, in rare cases, see that i can benefit from it and adopt it.
This video has transformed a lot unknown unknowns into known unknowns .
Great video as usual! Viktor really knows how to dive into tools from a practitioner's perspective. The only thing is, I'm not 100% on board with his take on Backstage. It's no secret that Viktor has his own unique flavor of opinions about Backstage as it stands. Can't wait to see what he thinks when the new declarative approach to plugin development hits GA!
Thanks, I would like to see this kind video every year. It's very helpful.
This is a third year in a row i made such video and the plan is to continue every January.
I've been waiting for the video in 2024. Thank you
For observability a scalable Prometheus option could be an OTEL Target Allocator and Grafana's MIMIR as a backend. Additional benefit OTEL TA also respects Prometheus Operators CRs! Which is helpful for service discovery!
Hey @powersurge5576, full disclosure here, I work for groundcover.
To your point, just wanted to mention that we use Victoria Metrics for the metrics backend which scales great for both short and long-term metric storage.
We also use oTel collector for ingesting logs and traces, however we use a VM agent for the metrics ingestion as it is more efficient and lightweight.
We have a really generous free tier - feel free to check us out. We’d love your feedback :)
Thank you for yet another excellent video. Without your advice I'd be lost.
I like your video and very helpful to me. , the one which I felt missing is Backup and recovery management in Kubernetes. It would be nice if you could also suggest a tool on that for 2024
I'm shocked for Database section there was not any mention of Flyway or Liquidbase. These do the necessary work of deploying SQL queries which is great. However, I do appreciate the mention of Atlas Operator. It's a cool operator that doesn't get enough mention.
Given that more and more apps are running in kubernetes it makes a lot of sense to bundle them with schemas in s way native to Kubernetes. That's why I prefer atlas operator. Flyway and Liquibase are great but did not yet, as far as I know, do the work of making a kubernetes CRD.
@@DevOpsToolkit Agreed! Would love to see more K8s Operators down the road. Bitnami's Sealed Secrets was mentioned and it has a special place for my team. Helps with our deployments of secrets in conjunction with AWS IRSA usage in AWS EKS. We use Bitnami Sealed Secrets as a backup just in case AWS STS has issues.
Amazing video! Just out of curiosity, how would you deploy the first kubernetes cluster that manages all the infrastructure?
That's the chicken or egg type of a problem. You can use a managed service like Upbound for control plane clusters. If you prefer a self-managed solution, there are two options that come to my mind. One is to start a local cluster with Crossplane and use it to create the first "real" control plane cluster. Once it's done, you can move Crossplane and that cluster definition there and, from there on, that "real" cluster would manage itself (apart from managing everything else). Alternatively, you can create that "real" cluster any other way (e.g., eksctl, hyperscaler console, etc.). Crossplane is primarily used to manage resources at scale so one of many clusters (not counting databases and everything else) is typically not an issue.
P.S. I assumed you asked about Crossplane. Please let me know if that's not the case and I'll adjust my answer.
@@DevOpsToolkit It was exactly that! Thanks a lot for the answer!
Thank you very much for amazing content as usual :)
Could you please elaborate on the infrastructure and service management using kubernetes ? Do you have a link or a video that you made about it ?
I understand that its a concept and I wondered if you have some practical examples of using it :)
I AM deeply involved with Crossplane (you'll find videos about it in this channel), but there are quite a few others as well.
Interesting, for security I feel like trivy-operator is the more complete product at the moment, although it is nice that kubescape is now part of CNCF
Hey victor, i know its been 9 months since this video and the landscape can change a lot in that time 😂😅 i was wondering if these tools are appropriate for a small team (5 devs) to manage self-managed servers as we operate in regions where there are NO cloud support
The landscape did not change much since then. Also, i will be making a new one around the beginning of the next year.
Unfortunately, I cannot day whether they are appropriate for such a small team. They are probably not. It might be an overkill to use kubernetes (many of those are related to kubernetes). I would go as far as to say that self-managed servers are not appropriate either. If there are only five devs in total you probably want to focus on developing whatever you're working on and not on managing infra and creating your own services. You should be using Google cloud run, azure Container Apps, fly.io, ir something similar. On the other hand, if you do have to self manage and if you do have a decently large scale (which you did not mention so i don't know whether that's the case), then kubernetes is the way to go.
Great video! Why didnt you include Elastic in observability? We have it in our company and it's a very powerful all-in-one solution now
There was no specific reason for not mentioning elastic other than trying to be brief. I also did not mention DataDog and Dynatrace and many others. I tried to mention only a top few, but that does not mean that there are no other good solutions (like Elastic).
You need a backup / mobility option for your managed and Kubernetes based databases…. Native operator doesn’t cut it if you have more than one database
That's true.
@@DevOpsToolkit we can chat in Paris, we have Kanister as a sandbox project now that interacts with those data services.
I missed this comment (and catching up with you in Paris). Sorry for that.
@@DevOpsToolkit let’s have a chat in the next place or sometime soon.
@90DaysOfDevOps send me a DM on Slack, Twitter, or LinkedIn and we'll organize a virtual coffee.
CNCF Incubating project Backstage is developing fast to overcome its complexity. Perhaps you'll revise your choice at BackstageCON next year?
Oh yeah. Backstage is moving so fast that everything i said can be invalidated in no time.
I hope your prediction for Gateway API comes true. Also you might missed Teller in the description
Thanks for letting me know. I'll add it to the description.
Thanks Very much Victor, very interesting video. Any chance we can have a one to one session?
Send me a DM on LinkedIn or Twitter.
I'd wary using kubernetes control plane for infrastructure. Heck, I even seen DNS servers on top of ESXi VMs, where the hosts relied on remote storage, which depended on DNS in the first place. A kubernetes cluster can die spectacularly in a real world environment, especially if it is provided by a cloud provider. To me infrastructure is critical enough not to rely on a platform with too many abstraction layers.
Awesome content and tools you shared 🎉. Cheers @DevOpsToolkit
Would be cool to make a video of tools to consider dropping in 2024 😅. One cant just add endlessly
I like the idea, even though that will get me into even more trouble.
How about this. A video near the end of a year with droppings, and a video around the beginning of a year with additions?
Hi,
I have 8 years of experience as a programmer, I have been working as DevOps for 2 years? What technologies do you recommend learning about? With AWS I have 4 certificates, with Kubernetes I have CKAD. Overall, my stack is mainly AWS, Linux, Gitlab CI/CD, bash, Terraform and Ansible to a lesser extent, plus python, but I do very little development in it.
I mainly use Zabbix for monitoring.
If you're already experienced with Kubernetes, you might want to explore Kubernetes-native tools like Prometheus, Argo CD, Crossplane, etc.
@@DevOpsToolkit I forgot, in the project I use Skaffold and ArgoCD. For monitoring Prometheus + Grafana , and Longhorn for storage, sometimes I will write a small HELM package, but I prefer someone else's :)
@pawecyrklaf4729 that's already an impressive portfolio, assuming that you're using those in production at a decent scale.
@@DevOpsToolkit The scale is not huge, so sometimes I have to go back to documentation, e.g. Ansible. As an AWS Community Builder, I have the opportunity to play around with more expensive services on AWS
Thanks for sharing the information, how about redhat advanced cloud security for kubernetes, can you please share your opinion, thanks
Unfortuantely, I have only a superficial experience with it so I can't comment on it or compare it with other solutions.
I would add Apisix under Netowrking/Ingress
Victor, could you please give an advice, how to create local development environments for Dev Teams? its a pain in the ass)) lets say, we need some Cloud services and DBs, containers accessible from the internet etc.
The answer to that question would be too big for a comment here. If you can be a bit more specific, I can answer here.
@@DevOpsToolkit i think, i will use local K8s (Kind) and DB instances and other services in cloud deployed by Terraform for each developer. Or maybe there are some tools like Localstack is better ? Hvala puno
@dmitriimrcat i think it's easier and better to use remote clusters.
Hello, I live in Korea and my job is DevOps engineer.
I'm enjoying your RUclips.
Can I summarize your RUclips translation into Korean and upload it to my blog?
It's not for profit. It's just my hobby.
I'm studying various new tools while watching your RUclips, and I want to share them with my close friends.
Furthermore, I would like to share it with IT engineers in Korea.
Will it be possible?
And if you don't mind...
Can I also post it on the blog of my company?
Sure. Go for it. The only thing I'll ask is to always provide a link to the original.
@@DevOpsToolkit Of course! Thank you.
Thanks for another great video, we really needed a list after a lot of videos! I agree with most of the choices you made, except for maybe databases (for postgres I'd evaluate crunchy Pg) and pipelines (I'd rather use gitlab for now)
The beauty of Dagger is that it does not replace pipelines. You still need something like GitLab to run it. I, for example, use GitHub Actions but inside them there is invocation of Dagger. That alone would not be enough for me to use it but then there is the ability to run the same locally as well. In other words, instead of putting Shell scripts inside pipelines, I put Dagger code.
Thanks for the update on this, I think I should have specified I'm using remote development everywhere I can. Among the other problems it solves it makes running pipelines feel like they were "local' as they are just like one of the other services. But I watched your video about it and still think dagger is a good idea for people who want to have the same behaviour in different environments
@IvanRizzante oh yeah. My preference is to use remote environments as well and in those cases dagger does not really provide much, if any value. Unfortinately, remote environments adoption is slower than i hoped and that's one of the predictions i had in the past and failed.
brilliant thanks Victor :)
Most Valuable Video™2024 ❤Thanks for sharing👍
No ingress picks for 2024? (=
The picks are mostly among those I invested heavilty in 2023. There's no Ingress that sparked my attention lately. Now it's mostly about Gate API implementations.
Ok, let's start over again
It would be better to separate out the ones which are open source
Good suggestion. I'll try to remember that the next time I create a similar video.
Hey dude.
In 2024?
Or better for ? 😂🎉
k8 is what happen when there is a lack of leadership on the project, and they simply accept PRs about any feature, in 2030 k8s will become an OS that you can install from an iso file and then use Microsoft Office
I want same but for data engineering