Project Harbor Community, 22 Feb 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF Harbor's Community Zoom Meeting - Feb 22, 2023

Description

CNCF Harbor's Community Zoom Meeting

A

All right, we're recording hello everyone. My name is julim vasilif and I'm. The community manager for Harbor today is February 22nd and as the official community meeting. So please follow the code of contact and just behave practically um with that I'm gonna paste once again, the community meeting notes into the chat. So if you can add yourself audio topics at your updates or whatever you want to discuss with us today. So with that I'm gonna share that same document with you.

A

A

Okay, I hope everyone can see that and with this one yeah I can see no much about topics at it uh and not many folks added themselves. But uh can you can you go ahead and then talk about the cloud events.

A

B

A

B

Hi hi Olly uh can I share my screen. Oh yeah, sure.

B

Okay, well guys uh I'm telling you um today. My topic is discussing the proposal of sports political event for webhook, um but actually it's not only for college Cloud event, but also we do some improvement to um to improve the better web hook management. So uh there are two major features for this proposal. One is webhook refactor and another is cloud event. Support.

B

um Firstly, let's quickly go through the background: uh Cloud events, uh it's a spec which just something like oci images back or Jose is built distribution spec, but this is for event so and also it's under the same there.

B

um So if Harbor can support to integrate the cloud event to the web hook, that can help better connect to the different vendors who follow the Cloud events back. That's that that is benefit for Harbor okay, so we have two goals which uh actually related to the Futures. One is refactor. The web hook code base to migrate the job to the common tasks framework. uh I will explain why we need to do this in the following section and another is uh to add the integration of cloud event.

B

uh Also, we have two known goals. One is one is we will not return all the web hook job histories, because in the previous Hardware versions, this table uh has never been cleaned up, so the data volume of this table may be held. So it's not a good chance to migrate all the old data, so we plan only to only migrate the last one for every event type and another long goal is we don't want to implement all combinations of cloud events back because it's all comes back.

B

The user have have enough blacks both to construct their data structure so, for example, they can use XML or Json or any other which we which follows the spec format.

B

How about will provide one standard format that is just implemented?

B

Next, let's move to the implementation for the front end. Is it's the UI for for this proposal. uh We will add the payload formats option selector in in the editor page in the web hook policy. Edit page, um it's not a required option and it's only only valued when the notific type is HTTP.

B

Why we not support slack, because we don't think slack is a note fee type and we plan to migrate the these two to a payload formats, uh so current So. Currently we only spot Cloud events for HTTP and in the future release we will migrate the slack as a payload format.

B

And The Following part has no change.

B

Okay, next is the page for the job histories in previous. There is no no such page for for for for the webhole policy, there is only a last trigger you can see. You can only sing the last Trigger Time for one event type, but but we refactor this part to the Excursion histories. uh Just like replication, Tech retention.

B

um There is a execution lists and when you click one execution, it will be redirected to the task page in a task page. You can see the task status, the task startime and time, and also you can click the logs button to see the logs when the job running.

B

Okay, this is the change for the UI part for the UI part and for the back end um regarding the refactor. We we can uh first thing I want to explain why we need to refactor this part. That is in the previous version. Hopper has unified the scheduler and task framework, but uh the webhook is a legacy vendor. It's it's implemented the task and Excursion by itself in a in a safe way, so but other jobs, vendors such as replication tank retention, GC using the unified way.

B

So we want to migrate the webhoot job to the unified way that can retrieve better management.

B

uh Regarding these parts, it can be summarized as following steps. Firstly, we need to create web hook, job by manager or controller provided by the task package. This package is, is our unified task framework and in the next we need to update the job implementation, uh uh because in the previous implementations there is no no debug logs for that, so user cannot, oh well what happened when the job running, so we will add more departments for this, and next is. We will introduce the new apis as a unified style for the operations of webhole jobs.

B

uh Actually, there are three apis. The first is list executions for a policy. The list API can relate it to. This can relate it to this page. When user click one policy, then it can display the executions under this policy that caught that call called this API these exclusions and the next is the tasks API.

B

uh This API can relate it to this page. That means, when you click one execution in this picture, and then it will call the API to list the tasks under the executions.

B

And next is uh last: is the get log API? This API can also relate it to uh this. This icon, when you, when you click the logs icon, then UI will call the log API to retrieve the logs for the job.

B

The job will looks like some information uh when, when it starts uh job and the what's the request body and the response code and successfully, then something like this foreign.

B

We also need to adjust the Legacy API Handler Logic for the compatible for the back of water compatible, because uh so so we we do not remove the old apis, but uh due to the refactor for the webhook job Manager Management, we need to do some of the just the adjustment, but actually the two apis. uh One is list jobs under the project and another is uh return. The last trigger information of the project weapon policy.

B

These two apis is not used wisely widely, so we will add the deprecated marker in the Swagger and remove them in the some future release.

B

Next is migrates the old job roles to to the new tables, um because in the previous uh in the previous implementations, the the web hook, job is dot stored in the table notification. Job and this table never be cleaned up. So so we need to uh migrate the data reading this table to the new table, execution and task, but we, but we will only migrate to the last job for every event, type uh from the performance View and the following Circle can help to do the migration.

B

So the basic logic for this circle is uh Specialists. We filter out the candidates, jobs and then do some conversion and finally insert into exclusion and task.

B

um By the way this decrypt can can can only help to migrate the data in the database. It's a static data, but if you have some runtime jobs, um you you need to cling this by the job service dashboard, because this these job running jobs, is in the radius queue so from the secret. We cannot clean this.

B

uh Finally, we need to drop the table notification job, but we, we may need to add the release notes to tell user. If he wants to keep the table. He should back up this table before upgrade.

B

Okay, next is the part about Cloud event. Cloud event is a spec for describing events data in a common way, so uh it defines some attributes and type and some Fields is required and some Fields is optional here. I list some important fields um first, the ID is, is a attribute to identify the events. It's the required attributes and the results is identify the context in which an event happened.

B

It's also a record and the spec version, spec or version is point out the which the events use the spec version right now, it's uh P1 the type is the field to describe the type of events related to the original occurrence.

B

And next for the following at Fields is optional one. So that means you can have these these fields or or don't have it's okay um to the content type if and stuff that the the the content type of the data value. This data schema is the schemer of the data. The data is the your event payload.

B

The subject is the uh describe the subject of your event, the term. That means uh what, when the event happened,.

B

Okay, next uh is the event type mapping that means in uh in previous versions. When you receive a webhold payload, you will got the event type for this event, and if you choose the cloud event, payload format, there is a event type um mapping to the old old type in the cloud event way. It's the style is the domain. Dot object, dot action, so you can refer to this table to kill the new type in the cloud event way.

B

Regarding the regarding the interface, we need to define a interface to handle the event data formation, because if we have this interface, the user from Community, who wants to contribute their custom or their wanted vendors, they can just add adding the driver just by Implement in implement the method implements. The interface um tell the following is just to demo uh case for how to implement the methods. For example, this is a Json formatter.

B

uh It implements the format method and it's do something: blah blah and finally, using Json to encode the payload and then return and then, however, will use this data uh send to the user defined endpoints.

B

And next. This is the cloud eventful method, so we'll do something blah blah and finally use cloud event way to encode the payload and then return the data.

B

Okay, uh finally, there are two examples to show the payload in the cloud event way. First is the push aspect um you can see. The spec version is 1.0 currently because we follow this back version, so so this value is fixed, uh but if, in the future we follow the to the 2.0, we should update list in in the in the value, and next is the type. The type is the event type you can following the uh this table, and next is the source source means.

B

um What's the who? Oh, that means Who is the source of the event for for for this case, uh this is a policy. The policies ID is wrong and under the projects one, so that means of the ID equals one policies. One web, one web Hood policy under the project. One um trigger this event- and the next is the id id- is the unique ID um for the hardware case. We will use the job ID to to field to fill in this value.

B

The time is, the event happened in in the work in what time and Operator Operator is a hardware application extension value. It's not in the cloud events back, but uh but it's the hardware application Fields, so uh it it's. It's represents the uh event operator of this of this event and at last year, is to to fill data content, type and data. The data is your payload data encoded by the data contents type way.

B

It's a for my case. It's my my content. Tab is application.json, so my data was encoded into Json format.

B

And next is the uh case for replication. It's something uh the the these fields is familiar with. The push artifact, uh only the only the difference is only the oh sorry, it's a typo uh this the data content tab should be application.json, and the data should be this one.

B

Okay, uh all right! That's that's all the part about this proposal. Any question.

C

So we have the uh sorry I have one question, so we have the ability to show the payload of uh the record set by Harbor right in uh yeah.

B

Yeah he can, he can click the logs to see the when, when the user click the logs, he can see the uh the logs and in our logs we print the webhook payload.

B

Okay, but we can also consider uh put the payload in one in the Excursion or task some column column.

C

Yeah I believe the it's not very convenience for user to collect several things to to get the payload and eventually get the people from the law right yeah. So.

B

C

In fact that we can have a color for them, yeah.

B

C

B

Just just like just like these picture shoes, uh we we can start eating the X3 attributes. Then user can click it to see the payload details.

D

C

Great. Thank you.

A

Thank you. Anyone else.

A

No, um can you I think with him uh proposed some changes on the namings on that that specific yeah.

B

Yeah I I see I, see that I will update.

C

Okay; okay, but by the way, the the Bible enhancement uh are also the support of cloud human is the anchor feature for Humber 2.8 on sunscreen thanks.

A

And one question from.

D

A

uh Is it a good idea to rework the web hook stuff all together with the cloud events, integration or.

A

I mean I'm asking shouldn't that be in like two different separate proposals, so we can have easy tracking of changes.

B

um Because they are, uh although they are not not conflict but um I, think we don't don't don't. We do not need to spill it. Two proposals I think we can come back track them by one.

A

And what what was the impact of changing the web hook stuff to for some other components? Is there any impact on this one I.

D

A

Need to change something else.

B

C

We just need to refactor the existing code base to to make it have ability to extend uh the the current support scope from uh slide to uh Cloud events and eventually in the future. We can easy to extend uh to add another can of a web hook a support. So so that's why we decided to uh do the refactor work for the laptop. So then, based on that, we can support uh Cloud event uh easily yeah.

A

And another one related to this one: uh do we have to if I'm, using the workbook in some way, is there a change to the way how I'm, using it and yeah.

C

A

C

Change and no API breaking yeah.

A

Okay, all right yeah, because we have to be careful just for the rest of the community. If, if someone is relying on this one yeah.

C

A

Okay, cool yeah that does everything from my side. There's questions.

A

Okay, uh thank you chenu. um The next one that we have on our agenda for today is give me a sec is Simon from VMware uh wants to briefly introduce do introduction of something called project. Narrows uh Simon. Are you with us.

E

F

G

G

G

uh Okay, time is the message.

A

Yeah we can see that yeah, okay,.

G

Sure so, uh hi uh um from uh Beijing um today I will be giving a very brief introduction to a hardware related open source project called project Narrows.

G

This project primarily focuses on enhancing security, and the initial idea actually came from the Harvard team and we collaborated with Chinese dimenso and other members from from the team to implement the first few versions of this project.

G

And uh project Narrows aims to address the security challenges in the coordinative space. We are, we we think that are currently. Organizations are typically Implement for qualitative security strategy.

G

To ensure security average includes our principles like shift security, left, continuous security controls or CSD pipeline Integrations, and something like accessibility, audit ability and availability and, as we all know, quality users, leverage Harbor to provide static analysis of vulnerabilities in images using scanners such as a tree and a clear the the static analysis of Scandal images after they have been pushed to the registry and project analysis, provides a unique addition to hardware, as it allows user to assess a security posture of kubernetes clusters at runtime.

G

This means image is asking at the time of inter introduction to a cluster, so vulnerabilities can uh so vulnerabilities are caught in your time and the images are flagged and workloads can be quarantined. So that's uh basically uh What uh project Narrows are provides.

G

And the two further assistance in the security of kubernetes infrastructure we designed and delivered a set of capabilities that enable users to emphasize the runtime security posture of kubernetes clusters and protect them from our vulnerabilities and attacks of these. Are these project errors?

G

Users can understand the overall security posture of the of their clusters by scanning the workloads either clusters and, at the same time, users can ensure that the actual security situations match their security, compliance expectations and a large any breaches user can uh Define some security policies or Security baselines in Project errors, and this with these policies uh it can quarantine, workloads, Source from a valuable images and stop the propagation of risks. Additionally, it's done kubernetes cluster. Additionally, with proton errors, the users cause service configurations following the Cris benchmark.

G

So uh so uh you know it's a GitHub report or this uh open source project and.

G

Here here is a basic workflow of this project. uh Firstly, image are cached in Harbor from uh any third party Registries maybe like to have. Then the image can be scanned in Harbor with the uh there's. Some scanners like TV, and the security data is generated in Hardware the security. Then the secure data can be consumed by Levels by some kubernetes clusters that have progenerals in store and then the scanning results of communist cluster is generated and and some are loss and some uh uh some some of us can be generated as well.

F

G

uh Let's watch our demo or progen error.

A

There's someone you stop sharing, so are we supposed to see some video or thank.

G

You yeah all right.

G

Thank you! That's great! No! Oh just just a.

F

Second, yeah.

A

F

D

E

A

G

A

H

Today, we're going to show a technical preview of our project, the environment we're using is set up with an upstream.

F

H

Today, we're going to show a technical preview of our project, the environment we're using is set up with an upstream kubernetes and is ready to use project Narrows under settings. A platform administrator must specify the security data source and today we're using Harbor in this setup. You'll create the secret to connect to Harbor and fill in the endpoint of the harbor instance. Then Define the image scanning interval.

H

After this setup, the images will be scanned periodically in Harbor to get the static image data which project Narrows will use to compare to when the workloads are running. You can also add known Registries to set replication rules in Harbor, and images from these Registries will be rotated within Harbor then scanned to get the security data on these images.

H

You also have a way to improve the performance by setting up an external redis. This will allow the security data from Harbor to be cached in redis and saved bandwidth.

H

Now that the configurations are complete, the security auditor can specify the scanning rules in the policy section where they can Define quarantining and security rules to create a policy. There are a number of fields to fill out, including how often the scan should run whether it's weekly, daily, hourly or custom users can also fill out the configuration for their open search or elasticsearch instances. So all the reports generated can be aggregated into the central place, they're using and easy to analyze.

H

In this section, the user can Define the baselines to set up the security expectations. This is important because workloads that violate any of these baselines to find will be flagged. This is also where you can add multiple, namespaces and workload labels to restrict the scope of the scanning.

H

Here you can choose whether you want to generate a report and quarantine. The workloads that are flagged then to view these reports you find them. In the assessment section. The application developer and security auditor are going to care most about this area. We have three types of reports generated correlating to the three cut scanners. We specified in the settings. So if you only choose one scanner, only one report will appear.

H

This visual shows the history of the number of containers scanned by project Narrows, we're going to dig into one of them to see more detailed information in this report. Default namespace was scanned, we'll drill down on the deploy test tube in this container, a critical vulnerability has been identified and because we chose to automatically quarantine workloads with issues, you can see that this is the action that was performed in this section. It shows the reports of cube Edge scanner that you specified in the policy Cube Edge checks, whether kubernetes is deployed securely.

H

Currently, we support two kinds of queue: bench checks: those are kubernetes policies and work on node security configurations, there's also risk scanning. The software packages inside the workload containers can be scanned by the risk scanner, and here are the scanning results. We also got the score from zero to five for measuring the severity of a vulnerability for each cve. Another area for the security auditor is to view not only the security posture, but also the risk Trends, which you organized into three categories: cluster namespace and workload. In the cluster section.

H

You can see how the winning workloads have been scanned and how many violate the Baseline in the namespace section, you can drill down by namespace to view the total workload scanned and how many are vulnerable to see the violation. In the workload section, you can see all the workloads that have been scanned by project Narrows, with the option to drill down hope you've enjoyed this Tech preview of project Narrows. In this video we've shown how you can add runtime scanning to your security Arsenal by creating policies, quarantining workloads and viewing your overall security posture.

F

Okay, so so that's it and.

F

G

We are, we are very excited to engage with our cognitive community and receive feedback and learn how people want to adopt or use these capabilities. So if any animal you guys are interested in now working with us more closely, but please email us, so the email is narrows at vmware.com.

G

Okay, that's it! Thank you for that.

A

Thank you. Simon I have few questions so practically you integrate with Harvard and you're reading the CV reports from.

E

A

Then consume them into the Narrows. Is that true yeah.

E

A

G

Yeah and uh actually uh we have a multiple security data source. The major uh the security data source is for Harbor from the CV DB. uh That's all that is for Hammer another. uh Another kind of uh secure data is something like the the rules of of misconfigurations. That is uh from uh another tactic. Yeah integrate.

A

With other tools like Cube scope, for example, yeah, because because I've seen, you used Cube bench to Overlay, evaluate the overall settings from the kubernetes cluster, which is good, but with the cube bench, you can do more in deep for, for the other deployments, for example, yeah.

G

Yeah right, so we uh in our roommate we plan to uh we plan to uh also integrate some uh some something uh kind of like a runtime uh security uh detection, something like a fair call or uh to to to get to collect the the runtime information or runtime uh or postures to project Narrows. So user can know much more clearly about the security yeah.

A

Yeah and last thing from my side uh how we can help from the community uh the hardware Community with that project: do you expect anything from my side or our site or uh just like to spread the word a bit that you integrate with? However,.

G

External capabilities uh for the runtime uh security detection by using project Narrows so because it uh integrates with integrated with Hardware very smoothly and so I think we can uh yeah. We can promote this project to uh to some of the hardware users.

F

A

Okay, uh I'll ask you to contact me offline with some with some blocks, or something like that, so we can share that with the community and they can give it a try or provide you some feedback. Okay, thank you. Thank you. So much yeah. Thank you all right. Thank you very much Simon anyone questions for Simon about this project or we can move on to the next topics.

A

Nope, okay, then, thank you, Simon. um Okay. Our next thing on our agenda is from Marcel uh right about.

I

Yes, yes, it's hopefully a short topic, but let's see um in Harbor operator, we right now have the problem. So the custom resource definitions are quite large in our operator and there have been like some minor additions in the recent pull requests and the problem has occurred now that the resources are too large for our Helm chart. So basically, Hub operator, together with the cids, is packaged at the hunt job and now that Helm chart exceeds one megabyte, which is by a Helm a specified maximum size for resources that can be deployed through hum.

I

So I was asking here if this has been a problem before or if there's already ideas on how to work around this, because, basically for Harbor operator development right now is basically difficult because I don't know the context about a lot of the options that are there and the CID and the crd makes by far the largest portion of size in the Hampshire right now.

I

If, if anyone has ideas, please speak up otherwise, I'm also happy to hear on slack um about ideas and how to go around this.

I

One thing could be: we split the helm shot into two charts or we remove unused options from the CID, but the problem there is I, don't know which options are used for the community, which makes it quite difficult to reduce the size of the overall chart.

A

What about home dependency.

I

We could try and add a dependency, but because the CID needs to be installed. First, it's not super easy to do. We could create a separate harm chart that we then installed as a dependency and try and order it correctly that it gets installed beforehand.

F

I

It's not very it's all very nice in my.

A

Experience yeah, it's not beautiful, yeah uh I I never came across this one in in the past, especially with Harbor anyone else from the maintenance team.

A

H

We can take that.

A

Yeah more deeper with the other uh like uh operate and maintain us. Maybe they can help just give them time. Maybe they can respond. Yeah.

I

Okay, I'll make another thread and have operator death, and maybe we find something yep cool.

E

C

You yeah I, I, yeah I know you want more uh maintenance from uh operator team to discuss the video from montages. Maybe some options can be removed from the Hampshire, but we should review them one by one and so carefully yeah, but so for your your problem. We can discuss in the sub Channel.

I

C

Could be a long story, so yeah.

I

Okay, thank you.

C

A

All right, um I can see some new faces, uh but uh before I give the floor for everyone else to introduce themselves, I'm, not sure. If we have introduced already omno uh to the team, I'm I'm. No can you can you introduce yourself? Please.

D

J

Don't know you yes, hello, everybody, my name is anmol and uh I am uh going to be the product manager for Harbor uh from VMware side, and, uh apart from that, I also look at packaging and deployment. So looking forward to working with everyone.

A

Welcome to the team.

G

Okay, yeah welcome yeah.

A

So, in short, uh unmo will take over from rotor and continue the the work on the Harbor project from VMware site. So if you have any roadmap in this kind of feature, requests and stuff- and more is your- is your guy um and I can see some some new faces?

A

C

I'm sorry I already, uh could you have uh have to add ammo into the uh Uncle Harbor uh repository? So then we can assign.

A

Yeah I'll do that.

C

So ammo, maybe you can share your GitHub handle so that we can add you into the Repository sure.

A

Thank you, uh I'll. Add you to the um to the GitHub repository uh organization level and our rdtd remaining lists. uh So you can. You can follow up on some stuff there. So thank you. Thank.

C

You thank you. Thank you.

A

All right uh the rest of the the folks on the call. uh Do you want to introduce yourself uh for for those of you who are feeling like it and new to the to the audience.

E

Yes, do you hear me yes,.

F

E

Do yeah hello, I'm, Thomas, I'm working for ovh company and we are using. We, we developed the arbor operator at start and we are using it to deploy the arbor registry of our customers.

A

Nice, um so you work together with the Jeremy or.

E

uh Yeah yeah, it's my manager.

A

Yeah, okay, great because uh in the past we we had Jeremy in Pierre. He was also in the ovh right.

E

Yeah and Simon too yeah yep, um yeah Jeremy is still here, but he's now. My manager and NPR left the company yeah two years ago. So.

A

Yeah he yeah, he went to data doc, I think um yeah yeah, and so what are the plans from ovh site you to to be become part of the maintenance team and represent the ovh into the Harbor community? Is that am I shooting right or.

E

Yep, that's it I.

C

E

Like to to help on our operator subject but like I say to to Yen, uh we are a bit late on the version uh our customers are on version Arbor 2.4, so we need to to upgrade to upgrade it. We are working on it and what we will we will be on last version. We will work on operator to to handle new Hubble versions.

A

Nice, okay, so we have uh since uh ovh was not super active in the last few months of say here uh yeah we have joined by uh Marcelo uh from transform and myself and with him on the other side, they're collaborating on.

H

A

Operator work, so I think uh you can get in touch with myself offline or something. So you can brief you out with the work and then what's going on Kurt right well, yeah.

E

Yeah we started to talk on the upload that old Dev channel on.

I

Selection: okay, yes,.

E

I

I also updated it today, because there's lots of problems with the CI and harbor operator right now,.

E

Yeah I saw it, you can yeah I'd continue to talk about it on slacky shots.

I

The only problem there is it takes really long because, like running CIA, every time takes like 30 minutes right now, so I'm just doing it when I have some time.

E

And I never played with it, so I can't I I can't help you very much for now. But yes, we we need, we need to fix it.

D

C

Hi Thomas I I have an idea so so cute is it possible that we can set up a weekly or by weekly meeting for a hub operator uh development uh to to discuss some engineering uh stuff like a CIA or PR issues? So uh it's I'd like to see uh several uh customer from Community or uh I'm interested on a hardware operator. So we can set up a regular engineer meetings to discuss uh engineering items so uh yeah. So is it possible to to organize that you can handle or.

I

A

Can set something up: yeah even I can do that through the CCF account. uh So.

I

A

I

Let's find a slot that works for everyone, but I would like to join and I can help setting it up and driving it or something yeah.

A

Yeah, the account is not my own right, it's a CNC F1. So whoever wants to to delete that therefore just hit record and then we can get it on uh online. So.

J

D

A

uh Okay, so I'll get in touch with you, so we can figure this out um uh all right. Thank you and welcome Thomas. um Thank you. Anyone else or we can.

A

I guess I suppose that's how you pronounce your name. We cannot hear you.

A

A

Yeah give it a try.

A

Nope I can see your move. Your lips moving, but yeah, okay, Marcel uh yeah I, have to drop this though I just want to bring in another topic. Okay, I guess: uh take your time. um I'm, I'm, gonna just drop another thing. uh We we had a request from the folks from uh what was their name.

A

uh Give me a sec, you something yeah, uvz I, suppose that's how you pronounce the name uh and they propose us to implement uh as part of our CI, to spin up a harbor instance uh for the testing of the pr. When the pr is closed, they can destroy that. So, if you can check I've added that into the uh into your community meetings, uh notes for today, I think it's pretty cool, I've tested it out with the uh for backstage and it's pretty responsive. So maybe we can think about.

A

uh They don't require anything from my side, our side just to let them do it, which is okay, I! Think. So, if you take a look at this and just provide some feedback, if it's you think it's okay for us to have such thing, but I think I find it nice.

A

If we do some stuff- and we can see the new instance of power up with our changes and just test them somewhere on the internet, not locally would be nice, so yeah.

C

It's a possible to to have someone from uh the same, to have a to give a timer for for Harbor to how to use that and I think it's useful for us. It's.

A

C

For the UI work, yeah yeah.

A

Yeah yeah, especially for the UI I'm yeah,.

D

A

All right I'll try to get in touch with them uh to to do. Actually we can drop a line into the pr that they've opened if they can join us on our next community meeting, so they can do some demo and elaborate a bit more yeah all right, okay, okay, thanks, okay and I. Guess second chance, man go ahead, no, no worries! Okay! We can try next time, eh. No, no, no, no worries!

A

um Okay! We have two minutes left part I. Think you've introduced yourself already right, yeah.

F

It's my second meeting.

A

Yeah, okay, all right, yes, are you are going to try again I suppose you're from the Linux Foundation main T program, if that's correct, yeah, just note, okay uh to brief you on this one, um we're gonna, do review of the mentees tomorrow, I hope, with uh oneion and and Vadim, and we're gonna select and continue with the program as expected, yeah, so yeah. Thank you very much again. Chenu. Thank you for uh for the demonstration Simon. Thank you as well. uh If that's it. Thank you very much.

A

Everyone I hope to see you again in two weeks and yeah have a nice rest of the day. Thank you.

E

Thank you. Thank you. Thank.

A